diff --git "a/measurement.json" "b/measurement.json" --- "a/measurement.json" +++ "b/measurement.json" @@ -2,7 +2,7 @@ "measurement": { "model.layers.0.self_attn": [ { - "accuracy": 0.9580099676784716, + "accuracy": 0.9579599970265439, "total_bits": 320921600, "q_proj": { "group_size": { @@ -66,7 +66,7 @@ } }, { - "accuracy": 0.9615060279243871, + "accuracy": 0.9614917635917664, "total_bits": 329244672, "q_proj": { "group_size": { @@ -130,7 +130,7 @@ } }, { - "accuracy": 0.9648278983015763, + "accuracy": 0.9647863538641679, "total_bits": 336188416, "q_proj": { "group_size": { @@ -194,7 +194,7 @@ } }, { - "accuracy": 0.9799904791932357, + "accuracy": 0.9799809299017254, "total_bits": 401721344, "q_proj": { "group_size": { @@ -258,7 +258,7 @@ } }, { - "accuracy": 0.9801941419902601, + "accuracy": 0.9801576247340754, "total_bits": 475443200, "q_proj": { "group_size": { @@ -322,7 +322,7 @@ } }, { - "accuracy": 0.9807896065084558, + "accuracy": 0.9807368218898773, "total_bits": 475642880, "q_proj": { "group_size": { @@ -386,7 +386,7 @@ } }, { - "accuracy": 0.9873390707530474, + "accuracy": 0.9873242072368923, "total_bits": 609923072, "q_proj": { "group_size": { @@ -438,7 +438,7 @@ } }, { - "accuracy": 0.9881942750591981, + "accuracy": 0.988159096554706, "total_bits": 610188288, "q_proj": { "group_size": { @@ -490,7 +490,7 @@ } }, { - "accuracy": 0.9891948574467709, + "accuracy": 0.989191010594368, "total_bits": 615184384, "q_proj": { "group_size": { @@ -542,7 +542,7 @@ } }, { - "accuracy": 0.9897961279279307, + "accuracy": 0.9897975654978501, "total_bits": 624115712, "q_proj": { "group_size": { @@ -594,7 +594,7 @@ } }, { - "accuracy": 0.9901267423441535, + "accuracy": 0.9901069151727777, "total_bits": 626637824, "q_proj": { "group_size": { @@ -658,7 +658,7 @@ } }, { - "accuracy": 0.9907404701960715, + "accuracy": 0.9907286669078627, "total_bits": 630519808, "q_proj": { "group_size": { @@ -722,7 +722,7 @@ } }, { - "accuracy": 0.991078717143912, + "accuracy": 0.9910670440447958, "total_bits": 637526016, "q_proj": { "group_size": { @@ -783,7 +783,7 @@ } }, { - "accuracy": 0.9918543312110399, + "accuracy": 0.991852257596819, "total_bits": 646987776, "q_proj": { "group_size": { @@ -844,7 +844,7 @@ } }, { - "accuracy": 0.9949971692342507, + "accuracy": 0.9949847871535703, "total_bits": 784904192, "q_proj": { "group_size": { @@ -905,7 +905,7 @@ } }, { - "accuracy": 0.9958498685767776, + "accuracy": 0.9958511351754791, "total_bits": 797982720, "q_proj": { "group_size": { @@ -966,7 +966,7 @@ } }, { - "accuracy": 0.9966111724313936, + "accuracy": 0.996609061172134, "total_bits": 911912960, "q_proj": { "group_size": { @@ -1018,7 +1018,7 @@ } }, { - "accuracy": 0.9976410932446781, + "accuracy": 0.9976443738529557, "total_bits": 942882816, "q_proj": { "group_size": { @@ -1070,7 +1070,7 @@ } }, { - "accuracy": 0.9990538940029708, + "accuracy": 0.9990543635856164, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -1124,7 +1124,7 @@ ], "model.layers.0.mlp": [ { - "accuracy": 0.9782323005952334, + "accuracy": 0.9782202494771857, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -1176,7 +1176,7 @@ } }, { - "accuracy": 0.9791372481145357, + "accuracy": 0.9791053643352107, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -1228,7 +1228,7 @@ } }, { - "accuracy": 0.9819421438794387, + "accuracy": 0.9819456495736775, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -1277,7 +1277,7 @@ } }, { - "accuracy": 0.9840327313071803, + "accuracy": 0.9840336457679146, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -1326,7 +1326,7 @@ } }, { - "accuracy": 0.9901343686015982, + "accuracy": 0.9901287312570372, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -1378,7 +1378,7 @@ } }, { - "accuracy": 0.9907936312650379, + "accuracy": 0.9907871445542887, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -1430,7 +1430,7 @@ } }, { - "accuracy": 0.9926116560634813, + "accuracy": 0.9926131030446604, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -1479,7 +1479,7 @@ } }, { - "accuracy": 0.9939957833603809, + "accuracy": 0.9940001721444883, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -1522,7 +1522,7 @@ } }, { - "accuracy": 0.9946407499281984, + "accuracy": 0.9946411553966371, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -1565,7 +1565,7 @@ } }, { - "accuracy": 0.995028724011622, + "accuracy": 0.9950304066664294, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -1617,7 +1617,7 @@ } }, { - "accuracy": 0.9956114633302939, + "accuracy": 0.995606797698297, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -1669,7 +1669,7 @@ } }, { - "accuracy": 0.997430244362668, + "accuracy": 0.9974309680493254, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -1721,7 +1721,7 @@ } }, { - "accuracy": 0.9977773904408279, + "accuracy": 0.997774467460419, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -1773,7 +1773,7 @@ } }, { - "accuracy": 0.9983576697934615, + "accuracy": 0.9983583417181906, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -1816,7 +1816,7 @@ } }, { - "accuracy": 0.9987043649154274, + "accuracy": 0.9987046939173811, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -1865,7 +1865,7 @@ } }, { - "accuracy": 0.9991657520203214, + "accuracy": 0.9991653028287386, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -1911,7 +1911,7 @@ } }, { - "accuracy": 0.9994252953482302, + "accuracy": 0.9994251264357253, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -1953,7 +1953,7 @@ ], "model.layers.1.self_attn": [ { - "accuracy": 0.9864757939388877, + "accuracy": 0.9864792321857653, "total_bits": 320921600, "q_proj": { "group_size": { @@ -2017,7 +2017,7 @@ } }, { - "accuracy": 0.9877178190570128, + "accuracy": 0.9877451681777051, "total_bits": 329244672, "q_proj": { "group_size": { @@ -2081,7 +2081,7 @@ } }, { - "accuracy": 0.9907136284991315, + "accuracy": 0.9907469961204027, "total_bits": 336188416, "q_proj": { "group_size": { @@ -2145,7 +2145,7 @@ } }, { - "accuracy": 0.993104424131544, + "accuracy": 0.9931729451606148, "total_bits": 401721344, "q_proj": { "group_size": { @@ -2209,7 +2209,7 @@ } }, { - "accuracy": 0.9932352484841096, + "accuracy": 0.9933124200293892, "total_bits": 475443200, "q_proj": { "group_size": { @@ -2273,7 +2273,7 @@ } }, { - "accuracy": 0.993879170794236, + "accuracy": 0.993870172453554, "total_bits": 475642880, "q_proj": { "group_size": { @@ -2337,7 +2337,7 @@ } }, { - "accuracy": 0.9934864899045542, + "accuracy": 0.99355722022684, "total_bits": 609923072, "q_proj": { "group_size": { @@ -2389,7 +2389,7 @@ } }, { - "accuracy": 0.9941575056628177, + "accuracy": 0.9941416923937044, "total_bits": 610188288, "q_proj": { "group_size": { @@ -2441,7 +2441,7 @@ } }, { - "accuracy": 0.9951102659106255, + "accuracy": 0.9950155847166714, "total_bits": 615184384, "q_proj": { "group_size": { @@ -2493,7 +2493,7 @@ } }, { - "accuracy": 0.995244678698088, + "accuracy": 0.9951392012207132, "total_bits": 624115712, "q_proj": { "group_size": { @@ -2545,7 +2545,7 @@ } }, { - "accuracy": 0.9969854156829809, + "accuracy": 0.9969725116695228, "total_bits": 626637824, "q_proj": { "group_size": { @@ -2609,7 +2609,7 @@ } }, { - "accuracy": 0.9974133950707159, + "accuracy": 0.9974157733744696, "total_bits": 630519808, "q_proj": { "group_size": { @@ -2673,7 +2673,7 @@ } }, { - "accuracy": 0.9970341558127027, + "accuracy": 0.9970228540662088, "total_bits": 637526016, "q_proj": { "group_size": { @@ -2734,7 +2734,7 @@ } }, { - "accuracy": 0.997508376444641, + "accuracy": 0.9975085501608095, "total_bits": 646987776, "q_proj": { "group_size": { @@ -2795,7 +2795,7 @@ } }, { - "accuracy": 0.998176209628582, + "accuracy": 0.9981939992622325, "total_bits": 784904192, "q_proj": { "group_size": { @@ -2856,7 +2856,7 @@ } }, { - "accuracy": 0.9986709901376775, + "accuracy": 0.9986674712088547, "total_bits": 797982720, "q_proj": { "group_size": { @@ -2917,7 +2917,7 @@ } }, { - "accuracy": 0.998234407290032, + "accuracy": 0.9982530158993445, "total_bits": 911912960, "q_proj": { "group_size": { @@ -2969,7 +2969,7 @@ } }, { - "accuracy": 0.9991800255681339, + "accuracy": 0.9991778942609304, "total_bits": 942882816, "q_proj": { "group_size": { @@ -3021,7 +3021,7 @@ } }, { - "accuracy": 0.9994745080016161, + "accuracy": 0.9994702832773328, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -3075,7 +3075,7 @@ ], "model.layers.1.mlp": [ { - "accuracy": 0.9787677824497223, + "accuracy": 0.9800858968182614, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -3127,7 +3127,7 @@ } }, { - "accuracy": 0.9783348629349157, + "accuracy": 0.9801951803659138, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -3179,7 +3179,7 @@ } }, { - "accuracy": 0.9827596545219421, + "accuracy": 0.9866483101719304, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -3228,7 +3228,7 @@ } }, { - "accuracy": 0.9827783123443001, + "accuracy": 0.9866749503110585, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -3277,7 +3277,7 @@ } }, { - "accuracy": 0.9916779571457913, + "accuracy": 0.9921675889115584, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -3329,7 +3329,7 @@ } }, { - "accuracy": 0.9952363579681045, + "accuracy": 0.9959408346759645, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -3381,7 +3381,7 @@ } }, { - "accuracy": 0.995260063362749, + "accuracy": 0.9959693519692672, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -3430,7 +3430,7 @@ } }, { - "accuracy": 0.9956984617992451, + "accuracy": 0.9953646126546358, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -3473,7 +3473,7 @@ } }, { - "accuracy": 0.9968835146803605, + "accuracy": 0.9966496840903634, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -3516,7 +3516,7 @@ } }, { - "accuracy": 0.9973897306542647, + "accuracy": 0.9975755538203215, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -3568,7 +3568,7 @@ } }, { - "accuracy": 0.9969309499781382, + "accuracy": 0.9976179364480471, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -3620,7 +3620,7 @@ } }, { - "accuracy": 0.997248651557847, + "accuracy": 0.9975226670503616, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -3672,7 +3672,7 @@ } }, { - "accuracy": 0.9974787329372606, + "accuracy": 0.9976930722202125, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -3724,7 +3724,7 @@ } }, { - "accuracy": 0.9973706099155702, + "accuracy": 0.9975672862247417, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -3767,7 +3767,7 @@ } }, { - "accuracy": 0.9973235453822111, + "accuracy": 0.9975260484375452, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -3816,7 +3816,7 @@ } }, { - "accuracy": 0.9973246831642953, + "accuracy": 0.9975273189576048, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -3862,7 +3862,7 @@ } }, { - "accuracy": 0.9972513263162813, + "accuracy": 0.9974400850110933, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -3904,7 +3904,7 @@ ], "model.layers.2.self_attn": [ { - "accuracy": 0.9981004162446449, + "accuracy": 0.9981189544655775, "total_bits": 320921600, "q_proj": { "group_size": { @@ -3968,7 +3968,7 @@ } }, { - "accuracy": 0.9982715690214383, + "accuracy": 0.9982857319869494, "total_bits": 329244672, "q_proj": { "group_size": { @@ -4032,7 +4032,7 @@ } }, { - "accuracy": 0.9985498411482886, + "accuracy": 0.9985529260807916, "total_bits": 336188416, "q_proj": { "group_size": { @@ -4096,7 +4096,7 @@ } }, { - "accuracy": 0.9989791929133629, + "accuracy": 0.9989791638953122, "total_bits": 401721344, "q_proj": { "group_size": { @@ -4160,7 +4160,7 @@ } }, { - "accuracy": 0.9990200108211291, + "accuracy": 0.9990197237777082, "total_bits": 475443200, "q_proj": { "group_size": { @@ -4224,7 +4224,7 @@ } }, { - "accuracy": 0.9990734019757885, + "accuracy": 0.9990638051378099, "total_bits": 475642880, "q_proj": { "group_size": { @@ -4288,7 +4288,7 @@ } }, { - "accuracy": 0.9991624246498472, + "accuracy": 0.9991609464938703, "total_bits": 609923072, "q_proj": { "group_size": { @@ -4340,7 +4340,7 @@ } }, { - "accuracy": 0.9992274578268591, + "accuracy": 0.9992157419359213, "total_bits": 610188288, "q_proj": { "group_size": { @@ -4392,7 +4392,7 @@ } }, { - "accuracy": 0.9992858364589905, + "accuracy": 0.9992871094298991, "total_bits": 615184384, "q_proj": { "group_size": { @@ -4444,7 +4444,7 @@ } }, { - "accuracy": 0.9992989457555508, + "accuracy": 0.9992991817722979, "total_bits": 624115712, "q_proj": { "group_size": { @@ -4496,7 +4496,7 @@ } }, { - "accuracy": 0.9995017176199901, + "accuracy": 0.9995037533931042, "total_bits": 626637824, "q_proj": { "group_size": { @@ -4560,7 +4560,7 @@ } }, { - "accuracy": 0.9995685118789736, + "accuracy": 0.9995715326482528, "total_bits": 630519808, "q_proj": { "group_size": { @@ -4624,7 +4624,7 @@ } }, { - "accuracy": 0.9995194366201758, + "accuracy": 0.9995223428367784, "total_bits": 637526016, "q_proj": { "group_size": { @@ -4685,7 +4685,7 @@ } }, { - "accuracy": 0.999593421522724, + "accuracy": 0.9995955619961023, "total_bits": 646987776, "q_proj": { "group_size": { @@ -4746,7 +4746,7 @@ } }, { - "accuracy": 0.9997149538493862, + "accuracy": 0.9997148106462861, "total_bits": 784904192, "q_proj": { "group_size": { @@ -4807,7 +4807,7 @@ } }, { - "accuracy": 0.9997818131784075, + "accuracy": 0.9997833367241057, "total_bits": 797982720, "q_proj": { "group_size": { @@ -4868,7 +4868,7 @@ } }, { - "accuracy": 0.9997456267623133, + "accuracy": 0.9997451584541092, "total_bits": 911912960, "q_proj": { "group_size": { @@ -4920,7 +4920,7 @@ } }, { - "accuracy": 0.9999069705617761, + "accuracy": 0.999907170116019, "total_bits": 942882816, "q_proj": { "group_size": { @@ -4972,7 +4972,7 @@ } }, { - "accuracy": 0.9999354342795223, + "accuracy": 0.9999354118113651, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -5026,7 +5026,7 @@ ], "model.layers.2.mlp": [ { - "accuracy": 0.998356753568116, + "accuracy": 0.9983650605733457, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -5078,7 +5078,7 @@ } }, { - "accuracy": 0.9983928095745412, + "accuracy": 0.9984011772627893, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -5130,7 +5130,7 @@ } }, { - "accuracy": 0.9986323600536898, + "accuracy": 0.9986389994033074, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -5179,7 +5179,7 @@ } }, { - "accuracy": 0.9987320309798968, + "accuracy": 0.9987379844839636, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -5228,7 +5228,7 @@ } }, { - "accuracy": 0.9993226272299102, + "accuracy": 0.9993271879655751, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -5280,7 +5280,7 @@ } }, { - "accuracy": 0.9993739144778565, + "accuracy": 0.999377539087283, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -5332,7 +5332,7 @@ } }, { - "accuracy": 0.999513078382925, + "accuracy": 0.9995156614305941, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -5381,7 +5381,7 @@ } }, { - "accuracy": 0.9996203480435437, + "accuracy": 0.9996229598151618, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -5424,7 +5424,7 @@ } }, { - "accuracy": 0.9996627930208648, + "accuracy": 0.9996644404324654, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -5467,7 +5467,7 @@ } }, { - "accuracy": 0.9996559167721946, + "accuracy": 0.9996581342757532, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -5519,7 +5519,7 @@ } }, { - "accuracy": 0.9996977955789158, + "accuracy": 0.9996994680136835, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -5571,7 +5571,7 @@ } }, { - "accuracy": 0.9998212374354664, + "accuracy": 0.99982232203413, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -5623,7 +5623,7 @@ } }, { - "accuracy": 0.9998465135423956, + "accuracy": 0.9998473078502637, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -5675,7 +5675,7 @@ } }, { - "accuracy": 0.9998939449433237, + "accuracy": 0.9998942665682223, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -5718,7 +5718,7 @@ } }, { - "accuracy": 0.999907999667094, + "accuracy": 0.9999082014761179, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -5767,7 +5767,7 @@ } }, { - "accuracy": 0.9999392223238063, + "accuracy": 0.9999399659603736, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -5813,7 +5813,7 @@ } }, { - "accuracy": 0.9999629726394472, + "accuracy": 0.9999626983067413, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -5855,7 +5855,7 @@ ], "model.layers.3.self_attn": [ { - "accuracy": 0.9976706416591218, + "accuracy": 0.9976771032731784, "total_bits": 320921600, "q_proj": { "group_size": { @@ -5919,7 +5919,7 @@ } }, { - "accuracy": 0.9979281858786156, + "accuracy": 0.9979369450164469, "total_bits": 329244672, "q_proj": { "group_size": { @@ -5983,7 +5983,7 @@ } }, { - "accuracy": 0.9981088430473679, + "accuracy": 0.9981150864378402, "total_bits": 336188416, "q_proj": { "group_size": { @@ -6047,7 +6047,7 @@ } }, { - "accuracy": 0.9987393644099173, + "accuracy": 0.9987459660164619, "total_bits": 401721344, "q_proj": { "group_size": { @@ -6111,7 +6111,7 @@ } }, { - "accuracy": 0.9987889832179797, + "accuracy": 0.9987947420266113, "total_bits": 475443200, "q_proj": { "group_size": { @@ -6175,7 +6175,7 @@ } }, { - "accuracy": 0.9988880921155214, + "accuracy": 0.9988931231200695, "total_bits": 475642880, "q_proj": { "group_size": { @@ -6239,7 +6239,7 @@ } }, { - "accuracy": 0.999040278164964, + "accuracy": 0.9990438478753755, "total_bits": 609923072, "q_proj": { "group_size": { @@ -6291,7 +6291,7 @@ } }, { - "accuracy": 0.9991714515184101, + "accuracy": 0.9991752392563381, "total_bits": 610188288, "q_proj": { "group_size": { @@ -6343,7 +6343,7 @@ } }, { - "accuracy": 0.9993186128374777, + "accuracy": 0.9993203137756178, "total_bits": 615184384, "q_proj": { "group_size": { @@ -6395,7 +6395,7 @@ } }, { - "accuracy": 0.9993434579748857, + "accuracy": 0.9993451946954194, "total_bits": 624115712, "q_proj": { "group_size": { @@ -6447,7 +6447,7 @@ } }, { - "accuracy": 0.999431787058711, + "accuracy": 0.9994340310069291, "total_bits": 626637824, "q_proj": { "group_size": { @@ -6511,7 +6511,7 @@ } }, { - "accuracy": 0.9995054952604207, + "accuracy": 0.9995066186805305, "total_bits": 630519808, "q_proj": { "group_size": { @@ -6575,7 +6575,7 @@ } }, { - "accuracy": 0.9994704618951992, + "accuracy": 0.9994729931319231, "total_bits": 637526016, "q_proj": { "group_size": { @@ -6636,7 +6636,7 @@ } }, { - "accuracy": 0.9995541301997084, + "accuracy": 0.9995553545457753, "total_bits": 646987776, "q_proj": { "group_size": { @@ -6697,7 +6697,7 @@ } }, { - "accuracy": 0.9996833990965235, + "accuracy": 0.999683542275115, "total_bits": 784904192, "q_proj": { "group_size": { @@ -6758,7 +6758,7 @@ } }, { - "accuracy": 0.9997753672007668, + "accuracy": 0.9997762445311406, "total_bits": 797982720, "q_proj": { "group_size": { @@ -6819,7 +6819,7 @@ } }, { - "accuracy": 0.9997366397445532, + "accuracy": 0.9997366851342744, "total_bits": 911912960, "q_proj": { "group_size": { @@ -6871,7 +6871,7 @@ } }, { - "accuracy": 0.99988699812246, + "accuracy": 0.9998872912194776, "total_bits": 942882816, "q_proj": { "group_size": { @@ -6923,7 +6923,7 @@ } }, { - "accuracy": 0.999930620371168, + "accuracy": 0.9999308354760471, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -6977,7 +6977,7 @@ ], "model.layers.3.mlp": [ { - "accuracy": 0.9980925007870323, + "accuracy": 0.9980988626025225, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -7029,7 +7029,7 @@ } }, { - "accuracy": 0.9981392239661593, + "accuracy": 0.998145300110704, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -7081,7 +7081,7 @@ } }, { - "accuracy": 0.9984963998982781, + "accuracy": 0.9985019579333695, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -7130,7 +7130,7 @@ } }, { - "accuracy": 0.9986372680256241, + "accuracy": 0.9986423866725281, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -7179,7 +7179,7 @@ } }, { - "accuracy": 0.9990728394569535, + "accuracy": 0.9990758893913344, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -7231,7 +7231,7 @@ } }, { - "accuracy": 0.9991448850027824, + "accuracy": 0.9991479928752309, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -7283,7 +7283,7 @@ } }, { - "accuracy": 0.9993068670461837, + "accuracy": 0.999309379656456, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -7332,7 +7332,7 @@ } }, { - "accuracy": 0.9995010450580403, + "accuracy": 0.9995026558049416, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -7375,7 +7375,7 @@ } }, { - "accuracy": 0.9995520969754771, + "accuracy": 0.999553660470012, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -7418,7 +7418,7 @@ } }, { - "accuracy": 0.9995306638118467, + "accuracy": 0.9995323530349293, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -7470,7 +7470,7 @@ } }, { - "accuracy": 0.9995891842746029, + "accuracy": 0.999590683801982, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -7522,7 +7522,7 @@ } }, { - "accuracy": 0.9997599827321736, + "accuracy": 0.9997608924137527, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -7574,7 +7574,7 @@ } }, { - "accuracy": 0.999794566195066, + "accuracy": 0.9997952844163305, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -7626,7 +7626,7 @@ } }, { - "accuracy": 0.9998654653808396, + "accuracy": 0.9998659226847323, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -7669,7 +7669,7 @@ } }, { - "accuracy": 0.9998780398286486, + "accuracy": 0.9998784973776262, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -7718,7 +7718,7 @@ } }, { - "accuracy": 0.9999126699495766, + "accuracy": 0.9999130221917049, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -7764,7 +7764,7 @@ } }, { - "accuracy": 0.9999602005608674, + "accuracy": 0.9999603386233119, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -7806,7 +7806,7 @@ ], "model.layers.4.self_attn": [ { - "accuracy": 0.9982055108013906, + "accuracy": 0.9982163659051845, "total_bits": 320921600, "q_proj": { "group_size": { @@ -7870,7 +7870,7 @@ } }, { - "accuracy": 0.9983980027272513, + "accuracy": 0.9984083803076493, "total_bits": 329244672, "q_proj": { "group_size": { @@ -7934,7 +7934,7 @@ } }, { - "accuracy": 0.9986138571249811, + "accuracy": 0.9986179806292057, "total_bits": 336188416, "q_proj": { "group_size": { @@ -7998,7 +7998,7 @@ } }, { - "accuracy": 0.9990606323668831, + "accuracy": 0.9990639180729264, "total_bits": 401721344, "q_proj": { "group_size": { @@ -8062,7 +8062,7 @@ } }, { - "accuracy": 0.9991368493555408, + "accuracy": 0.9991404128898131, "total_bits": 475443200, "q_proj": { "group_size": { @@ -8126,7 +8126,7 @@ } }, { - "accuracy": 0.9991835049402557, + "accuracy": 0.9991884366364071, "total_bits": 475642880, "q_proj": { "group_size": { @@ -8190,7 +8190,7 @@ } }, { - "accuracy": 0.9993266524060777, + "accuracy": 0.9993282789364457, "total_bits": 609923072, "q_proj": { "group_size": { @@ -8242,7 +8242,7 @@ } }, { - "accuracy": 0.9993900637093344, + "accuracy": 0.9993932106973309, "total_bits": 610188288, "q_proj": { "group_size": { @@ -8294,7 +8294,7 @@ } }, { - "accuracy": 0.9994606049731374, + "accuracy": 0.9994745622634104, "total_bits": 615184384, "q_proj": { "group_size": { @@ -8346,7 +8346,7 @@ } }, { - "accuracy": 0.9994801197102979, + "accuracy": 0.9994941568982444, "total_bits": 624115712, "q_proj": { "group_size": { @@ -8398,7 +8398,7 @@ } }, { - "accuracy": 0.9995725602401715, + "accuracy": 0.9995741076081207, "total_bits": 626637824, "q_proj": { "group_size": { @@ -8462,7 +8462,7 @@ } }, { - "accuracy": 0.9996106217762357, + "accuracy": 0.9996138829982987, "total_bits": 630519808, "q_proj": { "group_size": { @@ -8526,7 +8526,7 @@ } }, { - "accuracy": 0.9996014083245475, + "accuracy": 0.9996007694372613, "total_bits": 637526016, "q_proj": { "group_size": { @@ -8587,7 +8587,7 @@ } }, { - "accuracy": 0.9996436708851865, + "accuracy": 0.9996462895681983, "total_bits": 646987776, "q_proj": { "group_size": { @@ -8648,7 +8648,7 @@ } }, { - "accuracy": 0.9997835980581218, + "accuracy": 0.9997836086212805, "total_bits": 784904192, "q_proj": { "group_size": { @@ -8709,7 +8709,7 @@ } }, { - "accuracy": 0.9998252654056016, + "accuracy": 0.9998247605552407, "total_bits": 797982720, "q_proj": { "group_size": { @@ -8770,7 +8770,7 @@ } }, { - "accuracy": 0.9998277554803184, + "accuracy": 0.9998267009535706, "total_bits": 911912960, "q_proj": { "group_size": { @@ -8822,7 +8822,7 @@ } }, { - "accuracy": 0.9999147789785638, + "accuracy": 0.9999149824725464, "total_bits": 942882816, "q_proj": { "group_size": { @@ -8874,7 +8874,7 @@ } }, { - "accuracy": 0.9999481839140093, + "accuracy": 0.9999482723100013, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -8928,7 +8928,7 @@ ], "model.layers.4.mlp": [ { - "accuracy": 0.9970573049626852, + "accuracy": 0.9970672799176291, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -8980,7 +8980,7 @@ } }, { - "accuracy": 0.9971276605991941, + "accuracy": 0.9971380433753917, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -9032,7 +9032,7 @@ } }, { - "accuracy": 0.9976153748207971, + "accuracy": 0.9976249901871932, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -9081,7 +9081,7 @@ } }, { - "accuracy": 0.9977789282014495, + "accuracy": 0.9977882612300547, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -9130,7 +9130,7 @@ } }, { - "accuracy": 0.9986395281004278, + "accuracy": 0.9986441454997188, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -9182,7 +9182,7 @@ } }, { - "accuracy": 0.9987450488107769, + "accuracy": 0.9987489532091116, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -9234,7 +9234,7 @@ } }, { - "accuracy": 0.9989682510495186, + "accuracy": 0.9989715061689678, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -9283,7 +9283,7 @@ } }, { - "accuracy": 0.9992830061696862, + "accuracy": 0.9992854284906858, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -9326,7 +9326,7 @@ } }, { - "accuracy": 0.9993546670792919, + "accuracy": 0.9993566162394065, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -9369,7 +9369,7 @@ } }, { - "accuracy": 0.9993125285070977, + "accuracy": 0.9993146810876695, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -9421,7 +9421,7 @@ } }, { - "accuracy": 0.9993984025755995, + "accuracy": 0.9994002590446096, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -9473,7 +9473,7 @@ } }, { - "accuracy": 0.9996489884674942, + "accuracy": 0.9996500520868913, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -9525,7 +9525,7 @@ } }, { - "accuracy": 0.9996996726840734, + "accuracy": 0.9997005540093309, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -9577,7 +9577,7 @@ } }, { - "accuracy": 0.999806814630957, + "accuracy": 0.9998074060575547, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -9620,7 +9620,7 @@ } }, { - "accuracy": 0.9998211724144456, + "accuracy": 0.9998216967490551, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -9669,7 +9669,7 @@ } }, { - "accuracy": 0.9998676334752848, + "accuracy": 0.9998680295814809, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -9715,7 +9715,7 @@ } }, { - "accuracy": 0.9999445028370246, + "accuracy": 0.9999446734712508, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -9757,7 +9757,7 @@ ], "model.layers.5.self_attn": [ { - "accuracy": 0.9985228615175736, + "accuracy": 0.9985334613409481, "total_bits": 320921600, "q_proj": { "group_size": { @@ -9821,7 +9821,7 @@ } }, { - "accuracy": 0.998691028180091, + "accuracy": 0.9987060662947203, "total_bits": 329244672, "q_proj": { "group_size": { @@ -9885,7 +9885,7 @@ } }, { - "accuracy": 0.9989104737576685, + "accuracy": 0.9989171256556323, "total_bits": 336188416, "q_proj": { "group_size": { @@ -9949,7 +9949,7 @@ } }, { - "accuracy": 0.9992174049839377, + "accuracy": 0.9992193118424008, "total_bits": 401721344, "q_proj": { "group_size": { @@ -10013,7 +10013,7 @@ } }, { - "accuracy": 0.9993155969698962, + "accuracy": 0.9993166032394296, "total_bits": 475443200, "q_proj": { "group_size": { @@ -10077,7 +10077,7 @@ } }, { - "accuracy": 0.9993460593549045, + "accuracy": 0.9993477032372826, "total_bits": 475642880, "q_proj": { "group_size": { @@ -10141,7 +10141,7 @@ } }, { - "accuracy": 0.9994390922549524, + "accuracy": 0.9994396296770949, "total_bits": 609923072, "q_proj": { "group_size": { @@ -10193,7 +10193,7 @@ } }, { - "accuracy": 0.9994821109269795, + "accuracy": 0.9994813136168217, "total_bits": 610188288, "q_proj": { "group_size": { @@ -10245,7 +10245,7 @@ } }, { - "accuracy": 0.9995377887236444, + "accuracy": 0.9995390078739116, "total_bits": 615184384, "q_proj": { "group_size": { @@ -10297,7 +10297,7 @@ } }, { - "accuracy": 0.9995527812524846, + "accuracy": 0.9995519243376819, "total_bits": 624115712, "q_proj": { "group_size": { @@ -10349,7 +10349,7 @@ } }, { - "accuracy": 0.9996711134567464, + "accuracy": 0.9996716413450869, "total_bits": 626637824, "q_proj": { "group_size": { @@ -10413,7 +10413,7 @@ } }, { - "accuracy": 0.9997016938011113, + "accuracy": 0.9997049867616672, "total_bits": 630519808, "q_proj": { "group_size": { @@ -10477,7 +10477,7 @@ } }, { - "accuracy": 0.9996917372765509, + "accuracy": 0.9996933895653408, "total_bits": 637526016, "q_proj": { "group_size": { @@ -10538,7 +10538,7 @@ } }, { - "accuracy": 0.999727329729419, + "accuracy": 0.99972956730543, "total_bits": 646987776, "q_proj": { "group_size": { @@ -10599,7 +10599,7 @@ } }, { - "accuracy": 0.9998190037440509, + "accuracy": 0.9998200238363719, "total_bits": 784904192, "q_proj": { "group_size": { @@ -10660,7 +10660,7 @@ } }, { - "accuracy": 0.9998569685573641, + "accuracy": 0.9998586260419535, "total_bits": 797982720, "q_proj": { "group_size": { @@ -10721,7 +10721,7 @@ } }, { - "accuracy": 0.9998458892254061, + "accuracy": 0.9998467095980519, "total_bits": 911912960, "q_proj": { "group_size": { @@ -10773,7 +10773,7 @@ } }, { - "accuracy": 0.9999250223749856, + "accuracy": 0.9999251497149664, "total_bits": 942882816, "q_proj": { "group_size": { @@ -10825,7 +10825,7 @@ } }, { - "accuracy": 0.9999561013935722, + "accuracy": 0.9999563539321045, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -10879,7 +10879,7 @@ ], "model.layers.5.mlp": [ { - "accuracy": 0.9963030411224616, + "accuracy": 0.9963145405054092, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -10931,7 +10931,7 @@ } }, { - "accuracy": 0.9964151566750125, + "accuracy": 0.9964267584054094, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -10983,7 +10983,7 @@ } }, { - "accuracy": 0.9970314018428326, + "accuracy": 0.997041434833878, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -11032,7 +11032,7 @@ } }, { - "accuracy": 0.9972368961102084, + "accuracy": 0.9972463959926053, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -11081,7 +11081,7 @@ } }, { - "accuracy": 0.9982389331256089, + "accuracy": 0.9982452647466409, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -11133,7 +11133,7 @@ } }, { - "accuracy": 0.9983710069209337, + "accuracy": 0.9983768165111542, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -11185,7 +11185,7 @@ } }, { - "accuracy": 0.9986444521499308, + "accuracy": 0.998649966657946, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -11234,7 +11234,7 @@ } }, { - "accuracy": 0.9990753829479218, + "accuracy": 0.9990785477780982, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -11277,7 +11277,7 @@ } }, { - "accuracy": 0.9991647875622699, + "accuracy": 0.9991674438903206, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -11320,7 +11320,7 @@ } }, { - "accuracy": 0.9991112166133366, + "accuracy": 0.9991143127217105, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -11372,7 +11372,7 @@ } }, { - "accuracy": 0.9992208169577154, + "accuracy": 0.9992233729970298, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -11424,7 +11424,7 @@ } }, { - "accuracy": 0.9995473851694873, + "accuracy": 0.999549137330369, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -11476,7 +11476,7 @@ } }, { - "accuracy": 0.9996117815424345, + "accuracy": 0.9996131525963152, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -11528,7 +11528,7 @@ } }, { - "accuracy": 0.9997522303050286, + "accuracy": 0.9997530932734279, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -11571,7 +11571,7 @@ } }, { - "accuracy": 0.9997693085307745, + "accuracy": 0.9997703965116096, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -11620,7 +11620,7 @@ } }, { - "accuracy": 0.9998256121149385, + "accuracy": 0.9998265063439152, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -11666,7 +11666,7 @@ } }, { - "accuracy": 0.9999304216992306, + "accuracy": 0.9999306883454617, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -11708,7 +11708,7 @@ ], "model.layers.6.self_attn": [ { - "accuracy": 0.9981894585254946, + "accuracy": 0.9981930167659333, "total_bits": 320921600, "q_proj": { "group_size": { @@ -11772,7 +11772,7 @@ } }, { - "accuracy": 0.9983981165446734, + "accuracy": 0.998406649028, "total_bits": 329244672, "q_proj": { "group_size": { @@ -11836,7 +11836,7 @@ } }, { - "accuracy": 0.9985633389534134, + "accuracy": 0.9985652849274246, "total_bits": 336188416, "q_proj": { "group_size": { @@ -11900,7 +11900,7 @@ } }, { - "accuracy": 0.9990160290740038, + "accuracy": 0.999021405648244, "total_bits": 401721344, "q_proj": { "group_size": { @@ -11964,7 +11964,7 @@ } }, { - "accuracy": 0.9991266917633382, + "accuracy": 0.9991306372379002, "total_bits": 475443200, "q_proj": { "group_size": { @@ -12028,7 +12028,7 @@ } }, { - "accuracy": 0.9991783020821842, + "accuracy": 0.9991778092164743, "total_bits": 475642880, "q_proj": { "group_size": { @@ -12092,7 +12092,7 @@ } }, { - "accuracy": 0.9993526089054189, + "accuracy": 0.9993574708504113, "total_bits": 609923072, "q_proj": { "group_size": { @@ -12144,7 +12144,7 @@ } }, { - "accuracy": 0.9994221413508058, + "accuracy": 0.9994245266149703, "total_bits": 610188288, "q_proj": { "group_size": { @@ -12196,7 +12196,7 @@ } }, { - "accuracy": 0.9994997354224324, + "accuracy": 0.9995014196948001, "total_bits": 615184384, "q_proj": { "group_size": { @@ -12248,7 +12248,7 @@ } }, { - "accuracy": 0.999516489278329, + "accuracy": 0.99951780097265, "total_bits": 624115712, "q_proj": { "group_size": { @@ -12300,7 +12300,7 @@ } }, { - "accuracy": 0.9995832511370903, + "accuracy": 0.9995850114347903, "total_bits": 626637824, "q_proj": { "group_size": { @@ -12364,7 +12364,7 @@ } }, { - "accuracy": 0.9996195159558403, + "accuracy": 0.9996184254261223, "total_bits": 630519808, "q_proj": { "group_size": { @@ -12428,7 +12428,7 @@ } }, { - "accuracy": 0.9996211149631754, + "accuracy": 0.9996228753589094, "total_bits": 637526016, "q_proj": { "group_size": { @@ -12489,7 +12489,7 @@ } }, { - "accuracy": 0.9996619694130985, + "accuracy": 0.9996616917319203, "total_bits": 646987776, "q_proj": { "group_size": { @@ -12550,7 +12550,7 @@ } }, { - "accuracy": 0.9997788637533391, + "accuracy": 0.9997786019046447, "total_bits": 784904192, "q_proj": { "group_size": { @@ -12611,7 +12611,7 @@ } }, { - "accuracy": 0.9998278835004097, + "accuracy": 0.9998276611593994, "total_bits": 797982720, "q_proj": { "group_size": { @@ -12672,7 +12672,7 @@ } }, { - "accuracy": 0.999826463307009, + "accuracy": 0.9998258229001964, "total_bits": 911912960, "q_proj": { "group_size": { @@ -12724,7 +12724,7 @@ } }, { - "accuracy": 0.9999112692404244, + "accuracy": 0.9999111344621173, "total_bits": 942882816, "q_proj": { "group_size": { @@ -12776,7 +12776,7 @@ } }, { - "accuracy": 0.9999484092573097, + "accuracy": 0.9999485311318973, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -12830,7 +12830,7 @@ ], "model.layers.6.mlp": [ { - "accuracy": 0.9955537727004603, + "accuracy": 0.9955783396175033, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -12882,7 +12882,7 @@ } }, { - "accuracy": 0.9956953878465452, + "accuracy": 0.9957163910332479, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -12934,7 +12934,7 @@ } }, { - "accuracy": 0.9963960122121008, + "accuracy": 0.9964144339686946, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -12983,7 +12983,7 @@ } }, { - "accuracy": 0.9966242066339442, + "accuracy": 0.9966410994529724, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -13032,7 +13032,7 @@ } }, { - "accuracy": 0.9978904186895019, + "accuracy": 0.9979016206374294, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -13084,7 +13084,7 @@ } }, { - "accuracy": 0.9980473185056135, + "accuracy": 0.9980568258385909, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -13136,7 +13136,7 @@ } }, { - "accuracy": 0.9983581672177503, + "accuracy": 0.9983655987797599, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -13185,7 +13185,7 @@ } }, { - "accuracy": 0.9988992994553164, + "accuracy": 0.9989048025913929, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -13228,7 +13228,7 @@ } }, { - "accuracy": 0.9990030968267667, + "accuracy": 0.9990080943037021, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -13271,7 +13271,7 @@ } }, { - "accuracy": 0.9989372252633697, + "accuracy": 0.9989424018483413, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -13323,7 +13323,7 @@ } }, { - "accuracy": 0.9990670734918431, + "accuracy": 0.9990716827543158, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -13375,7 +13375,7 @@ } }, { - "accuracy": 0.9994595122376555, + "accuracy": 0.9994622971862555, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -13427,7 +13427,7 @@ } }, { - "accuracy": 0.9995356367802933, + "accuracy": 0.9995379316571512, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -13479,7 +13479,7 @@ } }, { - "accuracy": 0.999705543496499, + "accuracy": 0.9997070545183593, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -13522,7 +13522,7 @@ } }, { - "accuracy": 0.9997246665634999, + "accuracy": 0.9997260341371753, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -13571,7 +13571,7 @@ } }, { - "accuracy": 0.9997886756287986, + "accuracy": 0.9997897373855507, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -13617,7 +13617,7 @@ } }, { - "accuracy": 0.9999186018800461, + "accuracy": 0.9999189643360871, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -13659,7 +13659,7 @@ ], "model.layers.7.self_attn": [ { - "accuracy": 0.9976371014980894, + "accuracy": 0.9976417837958587, "total_bits": 320921600, "q_proj": { "group_size": { @@ -13723,7 +13723,7 @@ } }, { - "accuracy": 0.9979598188870832, + "accuracy": 0.9979639990549338, "total_bits": 329244672, "q_proj": { "group_size": { @@ -13787,7 +13787,7 @@ } }, { - "accuracy": 0.9981879319407438, + "accuracy": 0.9981890520767162, "total_bits": 336188416, "q_proj": { "group_size": { @@ -13851,7 +13851,7 @@ } }, { - "accuracy": 0.9987327964290192, + "accuracy": 0.998737685774502, "total_bits": 401721344, "q_proj": { "group_size": { @@ -13915,7 +13915,7 @@ } }, { - "accuracy": 0.9989152895776849, + "accuracy": 0.9989241218488467, "total_bits": 475443200, "q_proj": { "group_size": { @@ -13979,7 +13979,7 @@ } }, { - "accuracy": 0.9989461292953867, + "accuracy": 0.9989571549782628, "total_bits": 475642880, "q_proj": { "group_size": { @@ -14043,7 +14043,7 @@ } }, { - "accuracy": 0.9992026311198348, + "accuracy": 0.9992143563239982, "total_bits": 609923072, "q_proj": { "group_size": { @@ -14095,7 +14095,7 @@ } }, { - "accuracy": 0.9992470626572245, + "accuracy": 0.9992537113690847, "total_bits": 610188288, "q_proj": { "group_size": { @@ -14147,7 +14147,7 @@ } }, { - "accuracy": 0.999335052249463, + "accuracy": 0.9993374226124663, "total_bits": 615184384, "q_proj": { "group_size": { @@ -14199,7 +14199,7 @@ } }, { - "accuracy": 0.999363676056658, + "accuracy": 0.9993655870325471, "total_bits": 624115712, "q_proj": { "group_size": { @@ -14251,7 +14251,7 @@ } }, { - "accuracy": 0.9994481605918784, + "accuracy": 0.9994534195743894, "total_bits": 626637824, "q_proj": { "group_size": { @@ -14315,7 +14315,7 @@ } }, { - "accuracy": 0.9995105340586681, + "accuracy": 0.9995110950089598, "total_bits": 630519808, "q_proj": { "group_size": { @@ -14379,7 +14379,7 @@ } }, { - "accuracy": 0.9994998900709968, + "accuracy": 0.9994992163816565, "total_bits": 637526016, "q_proj": { "group_size": { @@ -14440,7 +14440,7 @@ } }, { - "accuracy": 0.9995675137462584, + "accuracy": 0.9995683805134735, "total_bits": 646987776, "q_proj": { "group_size": { @@ -14501,7 +14501,7 @@ } }, { - "accuracy": 0.9997265752110827, + "accuracy": 0.9997264613936606, "total_bits": 784904192, "q_proj": { "group_size": { @@ -14562,7 +14562,7 @@ } }, { - "accuracy": 0.9997810846880862, + "accuracy": 0.9997807921057469, "total_bits": 797982720, "q_proj": { "group_size": { @@ -14623,7 +14623,7 @@ } }, { - "accuracy": 0.9997833287833553, + "accuracy": 0.9997829608374128, "total_bits": 911912960, "q_proj": { "group_size": { @@ -14675,7 +14675,7 @@ } }, { - "accuracy": 0.999877946684137, + "accuracy": 0.9998784409345765, "total_bits": 942882816, "q_proj": { "group_size": { @@ -14727,7 +14727,7 @@ } }, { - "accuracy": 0.9999329551417184, + "accuracy": 0.9999330804536217, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -14781,7 +14781,7 @@ ], "model.layers.7.mlp": [ { - "accuracy": 0.994309265362589, + "accuracy": 0.9943400307705528, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -14833,7 +14833,7 @@ } }, { - "accuracy": 0.9944858194181794, + "accuracy": 0.9945160397573521, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -14885,7 +14885,7 @@ } }, { - "accuracy": 0.9954006981692816, + "accuracy": 0.9954254086080351, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -14934,7 +14934,7 @@ } }, { - "accuracy": 0.9956792145967484, + "accuracy": 0.9957028399956854, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -14983,7 +14983,7 @@ } }, { - "accuracy": 0.9972591660916805, + "accuracy": 0.9972741211715498, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -15035,7 +15035,7 @@ } }, { - "accuracy": 0.9974715056780138, + "accuracy": 0.9974857456982136, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -15087,7 +15087,7 @@ } }, { - "accuracy": 0.997859438979312, + "accuracy": 0.9978715961700991, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -15136,7 +15136,7 @@ } }, { - "accuracy": 0.9985801353070297, + "accuracy": 0.9985877104887837, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -15179,7 +15179,7 @@ } }, { - "accuracy": 0.9987099387339855, + "accuracy": 0.9987171878548045, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -15222,7 +15222,7 @@ } }, { - "accuracy": 0.9986122041744622, + "accuracy": 0.9986201584535209, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -15274,7 +15274,7 @@ } }, { - "accuracy": 0.9987855153648477, + "accuracy": 0.9987918393392312, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -15326,7 +15326,7 @@ } }, { - "accuracy": 0.9992922867991423, + "accuracy": 0.9992960851247373, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -15378,7 +15378,7 @@ } }, { - "accuracy": 0.9993948324240352, + "accuracy": 0.9993980866611788, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -15430,7 +15430,7 @@ } }, { - "accuracy": 0.9996164388905623, + "accuracy": 0.999618501917116, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -15473,7 +15473,7 @@ } }, { - "accuracy": 0.999637760565077, + "accuracy": 0.9996396318617228, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -15522,7 +15522,7 @@ } }, { - "accuracy": 0.9997174118057286, + "accuracy": 0.9997189752267379, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -15568,7 +15568,7 @@ } }, { - "accuracy": 0.9998943674451622, + "accuracy": 0.9998948609726014, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -15610,7 +15610,7 @@ ], "model.layers.8.self_attn": [ { - "accuracy": 0.9963710366895324, + "accuracy": 0.9963809396875533, "total_bits": 320921600, "q_proj": { "group_size": { @@ -15674,7 +15674,7 @@ } }, { - "accuracy": 0.9967739415404043, + "accuracy": 0.996784584694787, "total_bits": 329244672, "q_proj": { "group_size": { @@ -15738,7 +15738,7 @@ } }, { - "accuracy": 0.9970087661947075, + "accuracy": 0.9970168228212156, "total_bits": 336188416, "q_proj": { "group_size": { @@ -15802,7 +15802,7 @@ } }, { - "accuracy": 0.998180404501526, + "accuracy": 0.998186536721493, "total_bits": 401721344, "q_proj": { "group_size": { @@ -15866,7 +15866,7 @@ } }, { - "accuracy": 0.998302772641182, + "accuracy": 0.9983054638693207, "total_bits": 475443200, "q_proj": { "group_size": { @@ -15930,7 +15930,7 @@ } }, { - "accuracy": 0.9983301070567808, + "accuracy": 0.9983362175132099, "total_bits": 475642880, "q_proj": { "group_size": { @@ -15994,7 +15994,7 @@ } }, { - "accuracy": 0.9989128654920741, + "accuracy": 0.9989151892889487, "total_bits": 609923072, "q_proj": { "group_size": { @@ -16046,7 +16046,7 @@ } }, { - "accuracy": 0.9989604329396236, + "accuracy": 0.9989655323718724, "total_bits": 610188288, "q_proj": { "group_size": { @@ -16098,7 +16098,7 @@ } }, { - "accuracy": 0.9990363301415193, + "accuracy": 0.9990392641017312, "total_bits": 615184384, "q_proj": { "group_size": { @@ -16150,7 +16150,7 @@ } }, { - "accuracy": 0.9990848757718739, + "accuracy": 0.9990872459388093, "total_bits": 624115712, "q_proj": { "group_size": { @@ -16202,7 +16202,7 @@ } }, { - "accuracy": 0.999148923217466, + "accuracy": 0.9991496862157395, "total_bits": 626637824, "q_proj": { "group_size": { @@ -16266,7 +16266,7 @@ } }, { - "accuracy": 0.9992006941649475, + "accuracy": 0.9992029002230418, "total_bits": 630519808, "q_proj": { "group_size": { @@ -16330,7 +16330,7 @@ } }, { - "accuracy": 0.9992373697459698, + "accuracy": 0.9992375025329622, "total_bits": 637526016, "q_proj": { "group_size": { @@ -16391,7 +16391,7 @@ } }, { - "accuracy": 0.9993000087867442, + "accuracy": 0.9993014835605496, "total_bits": 646987776, "q_proj": { "group_size": { @@ -16452,7 +16452,7 @@ } }, { - "accuracy": 0.9995791925803611, + "accuracy": 0.9995793530129289, "total_bits": 784904192, "q_proj": { "group_size": { @@ -16513,7 +16513,7 @@ } }, { - "accuracy": 0.9996461415614345, + "accuracy": 0.9996463852738472, "total_bits": 797982720, "q_proj": { "group_size": { @@ -16574,7 +16574,7 @@ } }, { - "accuracy": 0.9997129599122625, + "accuracy": 0.9997126889219018, "total_bits": 911912960, "q_proj": { "group_size": { @@ -16626,7 +16626,7 @@ } }, { - "accuracy": 0.9998101148460257, + "accuracy": 0.999810345054261, "total_bits": 942882816, "q_proj": { "group_size": { @@ -16678,7 +16678,7 @@ } }, { - "accuracy": 0.9999197943773317, + "accuracy": 0.9999195321426285, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -16732,7 +16732,7 @@ ], "model.layers.8.mlp": [ { - "accuracy": 0.9932989870247088, + "accuracy": 0.9933409322249261, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -16784,7 +16784,7 @@ } }, { - "accuracy": 0.993485080568414, + "accuracy": 0.9935229295178464, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -16836,7 +16836,7 @@ } }, { - "accuracy": 0.9945187780417895, + "accuracy": 0.994551063368195, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -16885,7 +16885,7 @@ } }, { - "accuracy": 0.994840329022784, + "accuracy": 0.994870452504409, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -16934,7 +16934,7 @@ } }, { - "accuracy": 0.996782283837858, + "accuracy": 0.9968025503974212, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -16986,7 +16986,7 @@ } }, { - "accuracy": 0.9970389990822265, + "accuracy": 0.9970578380713337, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -17038,7 +17038,7 @@ } }, { - "accuracy": 0.9974890608144434, + "accuracy": 0.9975046595852626, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -17087,7 +17087,7 @@ } }, { - "accuracy": 0.9983310838671107, + "accuracy": 0.9983407911893568, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -17130,7 +17130,7 @@ } }, { - "accuracy": 0.9984884075820446, + "accuracy": 0.9984978177633724, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -17173,7 +17173,7 @@ } }, { - "accuracy": 0.9983670586033871, + "accuracy": 0.9983763890831094, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -17225,7 +17225,7 @@ } }, { - "accuracy": 0.9985737441793868, + "accuracy": 0.9985827862432128, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -17277,7 +17277,7 @@ } }, { - "accuracy": 0.9991643720943677, + "accuracy": 0.9991691868360105, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -17329,7 +17329,7 @@ } }, { - "accuracy": 0.9992877157208951, + "accuracy": 0.999292261114246, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -17381,7 +17381,7 @@ } }, { - "accuracy": 0.9995463714003563, + "accuracy": 0.999548940086051, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -17424,7 +17424,7 @@ } }, { - "accuracy": 0.9995724283845017, + "accuracy": 0.9995748253637239, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -17473,7 +17473,7 @@ } }, { - "accuracy": 0.9996663753531481, + "accuracy": 0.9996682792215755, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -17519,7 +17519,7 @@ } }, { - "accuracy": 0.9998744614177236, + "accuracy": 0.9998751884497898, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -17561,7 +17561,7 @@ ], "model.layers.9.self_attn": [ { - "accuracy": 0.9952032213148317, + "accuracy": 0.9952150810705988, "total_bits": 320921600, "q_proj": { "group_size": { @@ -17625,7 +17625,7 @@ } }, { - "accuracy": 0.995742816281946, + "accuracy": 0.9957414457672521, "total_bits": 329244672, "q_proj": { "group_size": { @@ -17689,7 +17689,7 @@ } }, { - "accuracy": 0.9961313134745547, + "accuracy": 0.9961430269636606, "total_bits": 336188416, "q_proj": { "group_size": { @@ -17753,7 +17753,7 @@ } }, { - "accuracy": 0.997373745629662, + "accuracy": 0.9973879168300253, "total_bits": 401721344, "q_proj": { "group_size": { @@ -17817,7 +17817,7 @@ } }, { - "accuracy": 0.9977440294858656, + "accuracy": 0.9977583549916744, "total_bits": 475443200, "q_proj": { "group_size": { @@ -17881,7 +17881,7 @@ } }, { - "accuracy": 0.9978014685605702, + "accuracy": 0.9978112123514477, "total_bits": 475642880, "q_proj": { "group_size": { @@ -17945,7 +17945,7 @@ } }, { - "accuracy": 0.9983543067386276, + "accuracy": 0.9983692347611252, "total_bits": 609923072, "q_proj": { "group_size": { @@ -17997,7 +17997,7 @@ } }, { - "accuracy": 0.9984357990324497, + "accuracy": 0.9984489758744052, "total_bits": 610188288, "q_proj": { "group_size": { @@ -18049,7 +18049,7 @@ } }, { - "accuracy": 0.9985562622743217, + "accuracy": 0.9985605716509255, "total_bits": 615184384, "q_proj": { "group_size": { @@ -18101,7 +18101,7 @@ } }, { - "accuracy": 0.9986283744832403, + "accuracy": 0.9986371669526163, "total_bits": 624115712, "q_proj": { "group_size": { @@ -18153,7 +18153,7 @@ } }, { - "accuracy": 0.9988705451159101, + "accuracy": 0.9988761104055142, "total_bits": 626637824, "q_proj": { "group_size": { @@ -18217,7 +18217,7 @@ } }, { - "accuracy": 0.9989481117380293, + "accuracy": 0.9989531645061154, "total_bits": 630519808, "q_proj": { "group_size": { @@ -18281,7 +18281,7 @@ } }, { - "accuracy": 0.9989711233463726, + "accuracy": 0.9989750313719636, "total_bits": 637526016, "q_proj": { "group_size": { @@ -18342,7 +18342,7 @@ } }, { - "accuracy": 0.9990578076164973, + "accuracy": 0.9990635139769629, "total_bits": 646987776, "q_proj": { "group_size": { @@ -18403,7 +18403,7 @@ } }, { - "accuracy": 0.9994268703617548, + "accuracy": 0.9994296023720189, "total_bits": 784904192, "q_proj": { "group_size": { @@ -18464,7 +18464,7 @@ } }, { - "accuracy": 0.999523315774767, + "accuracy": 0.9995257061366972, "total_bits": 797982720, "q_proj": { "group_size": { @@ -18525,7 +18525,7 @@ } }, { - "accuracy": 0.9995626942009518, + "accuracy": 0.9995651525984469, "total_bits": 911912960, "q_proj": { "group_size": { @@ -18577,7 +18577,7 @@ } }, { - "accuracy": 0.9997272590959543, + "accuracy": 0.9997287151942912, "total_bits": 942882816, "q_proj": { "group_size": { @@ -18629,7 +18629,7 @@ } }, { - "accuracy": 0.9998713695248099, + "accuracy": 0.9998722708367399, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -18683,7 +18683,7 @@ ], "model.layers.9.mlp": [ { - "accuracy": 0.9929492355961549, + "accuracy": 0.992983805505853, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -18735,7 +18735,7 @@ } }, { - "accuracy": 0.9931369682675913, + "accuracy": 0.9931731326015372, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -18787,7 +18787,7 @@ } }, { - "accuracy": 0.9942131669897782, + "accuracy": 0.9942438712245539, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -18836,7 +18836,7 @@ } }, { - "accuracy": 0.9945557231181547, + "accuracy": 0.9945854164267841, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -18885,7 +18885,7 @@ } }, { - "accuracy": 0.9965789239657553, + "accuracy": 0.9965968394750043, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -18937,7 +18937,7 @@ } }, { - "accuracy": 0.9968464480418908, + "accuracy": 0.9968633459586846, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -18989,7 +18989,7 @@ } }, { - "accuracy": 0.9973193559991685, + "accuracy": 0.9973342834334624, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -19038,7 +19038,7 @@ } }, { - "accuracy": 0.9982319511473179, + "accuracy": 0.9982413878958476, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -19081,7 +19081,7 @@ } }, { - "accuracy": 0.9983970944426561, + "accuracy": 0.9984054371322456, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -19124,7 +19124,7 @@ } }, { - "accuracy": 0.9982680109770674, + "accuracy": 0.9982770436296338, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -19176,7 +19176,7 @@ } }, { - "accuracy": 0.9984850095290887, + "accuracy": 0.9984928012678498, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -19228,7 +19228,7 @@ } }, { - "accuracy": 0.9991162821258369, + "accuracy": 0.9991205429756328, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -19280,7 +19280,7 @@ } }, { - "accuracy": 0.9992441550191296, + "accuracy": 0.9992481284333687, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -19332,7 +19332,7 @@ } }, { - "accuracy": 0.9995221383869648, + "accuracy": 0.9995245146810224, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -19375,7 +19375,7 @@ } }, { - "accuracy": 0.9995487808789077, + "accuracy": 0.9995511829558956, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -19424,7 +19424,7 @@ } }, { - "accuracy": 0.9996447746985053, + "accuracy": 0.9996467883159456, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -19470,7 +19470,7 @@ } }, { - "accuracy": 0.9998687865506661, + "accuracy": 0.9998694574950557, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -19512,7 +19512,7 @@ ], "model.layers.10.self_attn": [ { - "accuracy": 0.9945719285230887, + "accuracy": 0.9946049569468749, "total_bits": 320921600, "q_proj": { "group_size": { @@ -19576,7 +19576,7 @@ } }, { - "accuracy": 0.9951097431935763, + "accuracy": 0.9951287064897386, "total_bits": 329244672, "q_proj": { "group_size": { @@ -19640,7 +19640,7 @@ } }, { - "accuracy": 0.9955582999085125, + "accuracy": 0.9955721364209527, "total_bits": 336188416, "q_proj": { "group_size": { @@ -19704,7 +19704,7 @@ } }, { - "accuracy": 0.9971086665203697, + "accuracy": 0.9971172613532919, "total_bits": 401721344, "q_proj": { "group_size": { @@ -19768,7 +19768,7 @@ } }, { - "accuracy": 0.9973965153882378, + "accuracy": 0.9974066324924168, "total_bits": 475443200, "q_proj": { "group_size": { @@ -19832,7 +19832,7 @@ } }, { - "accuracy": 0.9974740547569174, + "accuracy": 0.9974858372619277, "total_bits": 475642880, "q_proj": { "group_size": { @@ -19896,7 +19896,7 @@ } }, { - "accuracy": 0.9982322646599067, + "accuracy": 0.9982353680227932, "total_bits": 609923072, "q_proj": { "group_size": { @@ -19948,7 +19948,7 @@ } }, { - "accuracy": 0.9983497146320971, + "accuracy": 0.9983553549176768, "total_bits": 610188288, "q_proj": { "group_size": { @@ -20000,7 +20000,7 @@ } }, { - "accuracy": 0.998509888782313, + "accuracy": 0.9985130609650361, "total_bits": 615184384, "q_proj": { "group_size": { @@ -20052,7 +20052,7 @@ } }, { - "accuracy": 0.9985758123037062, + "accuracy": 0.9985807193933349, "total_bits": 624115712, "q_proj": { "group_size": { @@ -20104,7 +20104,7 @@ } }, { - "accuracy": 0.998707827376692, + "accuracy": 0.9987074684743819, "total_bits": 626637824, "q_proj": { "group_size": { @@ -20168,7 +20168,7 @@ } }, { - "accuracy": 0.9988107054837441, + "accuracy": 0.9988169522073708, "total_bits": 630519808, "q_proj": { "group_size": { @@ -20232,7 +20232,7 @@ } }, { - "accuracy": 0.9988429716935283, + "accuracy": 0.9988458134037884, "total_bits": 637526016, "q_proj": { "group_size": { @@ -20293,7 +20293,7 @@ } }, { - "accuracy": 0.9989588417504963, + "accuracy": 0.9989638924598694, "total_bits": 646987776, "q_proj": { "group_size": { @@ -20354,7 +20354,7 @@ } }, { - "accuracy": 0.9993603748612498, + "accuracy": 0.9993629280085626, "total_bits": 784904192, "q_proj": { "group_size": { @@ -20415,7 +20415,7 @@ } }, { - "accuracy": 0.9994778076774979, + "accuracy": 0.9994798760959193, "total_bits": 797982720, "q_proj": { "group_size": { @@ -20476,7 +20476,7 @@ } }, { - "accuracy": 0.9995339225095353, + "accuracy": 0.9995366309916502, "total_bits": 911912960, "q_proj": { "group_size": { @@ -20528,7 +20528,7 @@ } }, { - "accuracy": 0.9997189519926906, + "accuracy": 0.9997204689701137, "total_bits": 942882816, "q_proj": { "group_size": { @@ -20580,7 +20580,7 @@ } }, { - "accuracy": 0.9998669504482103, + "accuracy": 0.9998677943735138, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -20634,7 +20634,7 @@ ], "model.layers.10.mlp": [ { - "accuracy": 0.991751522610062, + "accuracy": 0.9917823280158796, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -20686,7 +20686,7 @@ } }, { - "accuracy": 0.9919772038334295, + "accuracy": 0.9920083645143007, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -20738,7 +20738,7 @@ } }, { - "accuracy": 0.993322126175228, + "accuracy": 0.9933486008330396, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -20787,7 +20787,7 @@ } }, { - "accuracy": 0.9937431831893168, + "accuracy": 0.993768384582118, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -20836,7 +20836,7 @@ } }, { - "accuracy": 0.995985034657152, + "accuracy": 0.9960024199987713, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -20888,7 +20888,7 @@ } }, { - "accuracy": 0.9962962222726721, + "accuracy": 0.9963106179707929, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -20940,7 +20940,7 @@ } }, { - "accuracy": 0.9968842428765798, + "accuracy": 0.9968959375431663, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -20989,7 +20989,7 @@ } }, { - "accuracy": 0.9979238388569731, + "accuracy": 0.9979315009948454, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -21032,7 +21032,7 @@ } }, { - "accuracy": 0.9981175057197872, + "accuracy": 0.9981243071195326, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -21075,7 +21075,7 @@ } }, { - "accuracy": 0.9979675925875965, + "accuracy": 0.9979758899854986, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -21127,7 +21127,7 @@ } }, { - "accuracy": 0.9982199917890524, + "accuracy": 0.9982264535991769, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -21179,7 +21179,7 @@ } }, { - "accuracy": 0.998963626101613, + "accuracy": 0.9989681051749932, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -21231,7 +21231,7 @@ } }, { - "accuracy": 0.999112726336247, + "accuracy": 0.9991157767607978, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -21283,7 +21283,7 @@ } }, { - "accuracy": 0.9994408546113654, + "accuracy": 0.9994432607567624, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -21326,7 +21326,7 @@ } }, { - "accuracy": 0.9994706893339753, + "accuracy": 0.999472840051902, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -21375,7 +21375,7 @@ } }, { - "accuracy": 0.999590139738039, + "accuracy": 0.999592144017745, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -21421,7 +21421,7 @@ } }, { - "accuracy": 0.9998462774766315, + "accuracy": 0.9998468524090162, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -21463,7 +21463,7 @@ ], "model.layers.11.self_attn": [ { - "accuracy": 0.9943421506568005, + "accuracy": 0.9943530567382511, "total_bits": 320921600, "q_proj": { "group_size": { @@ -21527,7 +21527,7 @@ } }, { - "accuracy": 0.9948741158372477, + "accuracy": 0.9948815766133761, "total_bits": 329244672, "q_proj": { "group_size": { @@ -21591,7 +21591,7 @@ } }, { - "accuracy": 0.9953428448030823, + "accuracy": 0.9953558158717657, "total_bits": 336188416, "q_proj": { "group_size": { @@ -21655,7 +21655,7 @@ } }, { - "accuracy": 0.9968719653000957, + "accuracy": 0.9968842818940941, "total_bits": 401721344, "q_proj": { "group_size": { @@ -21719,7 +21719,7 @@ } }, { - "accuracy": 0.9972916347415823, + "accuracy": 0.9973077470142591, "total_bits": 475443200, "q_proj": { "group_size": { @@ -21783,7 +21783,7 @@ } }, { - "accuracy": 0.9973474682161683, + "accuracy": 0.9973560659901092, "total_bits": 475642880, "q_proj": { "group_size": { @@ -21847,7 +21847,7 @@ } }, { - "accuracy": 0.9981763647182992, + "accuracy": 0.9981854550148311, "total_bits": 609923072, "q_proj": { "group_size": { @@ -21899,7 +21899,7 @@ } }, { - "accuracy": 0.9982454208166975, + "accuracy": 0.9982530208010423, "total_bits": 610188288, "q_proj": { "group_size": { @@ -21951,7 +21951,7 @@ } }, { - "accuracy": 0.9984206003382018, + "accuracy": 0.99842627032807, "total_bits": 615184384, "q_proj": { "group_size": { @@ -22003,7 +22003,7 @@ } }, { - "accuracy": 0.9984894428206118, + "accuracy": 0.9984963259806758, "total_bits": 624115712, "q_proj": { "group_size": { @@ -22055,7 +22055,7 @@ } }, { - "accuracy": 0.9986614431009481, + "accuracy": 0.9986661209871894, "total_bits": 626637824, "q_proj": { "group_size": { @@ -22119,7 +22119,7 @@ } }, { - "accuracy": 0.9987624427793842, + "accuracy": 0.9987659174948931, "total_bits": 630519808, "q_proj": { "group_size": { @@ -22183,7 +22183,7 @@ } }, { - "accuracy": 0.9988045462064052, + "accuracy": 0.998809299872894, "total_bits": 637526016, "q_proj": { "group_size": { @@ -22244,7 +22244,7 @@ } }, { - "accuracy": 0.9989192496593061, + "accuracy": 0.9989228944637274, "total_bits": 646987776, "q_proj": { "group_size": { @@ -22305,7 +22305,7 @@ } }, { - "accuracy": 0.9993288861587644, + "accuracy": 0.9993311454002795, "total_bits": 784904192, "q_proj": { "group_size": { @@ -22366,7 +22366,7 @@ } }, { - "accuracy": 0.9994476477762586, + "accuracy": 0.9994498426094651, "total_bits": 797982720, "q_proj": { "group_size": { @@ -22427,7 +22427,7 @@ } }, { - "accuracy": 0.9995124229279003, + "accuracy": 0.9995142588097798, "total_bits": 911912960, "q_proj": { "group_size": { @@ -22479,7 +22479,7 @@ } }, { - "accuracy": 0.9997070378280783, + "accuracy": 0.9997079629009884, "total_bits": 942882816, "q_proj": { "group_size": { @@ -22531,7 +22531,7 @@ } }, { - "accuracy": 0.9998642380447372, + "accuracy": 0.9998645180787303, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -22585,7 +22585,7 @@ ], "model.layers.11.mlp": [ { - "accuracy": 0.9906321426755503, + "accuracy": 0.9906661243815171, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -22637,7 +22637,7 @@ } }, { - "accuracy": 0.9908883297129681, + "accuracy": 0.9909236697774184, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -22689,7 +22689,7 @@ } }, { - "accuracy": 0.9924226858113941, + "accuracy": 0.9924532512300893, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -22738,7 +22738,7 @@ } }, { - "accuracy": 0.9929013503225226, + "accuracy": 0.9929308475632417, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -22787,7 +22787,7 @@ } }, { - "accuracy": 0.995415350716365, + "accuracy": 0.9954314702435544, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -22839,7 +22839,7 @@ } }, { - "accuracy": 0.9957787167084845, + "accuracy": 0.9957944927246947, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -22891,7 +22891,7 @@ } }, { - "accuracy": 0.9964476553233046, + "accuracy": 0.9964611581281612, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -22940,7 +22940,7 @@ } }, { - "accuracy": 0.9976210207923463, + "accuracy": 0.9976289805613066, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -22983,7 +22983,7 @@ } }, { - "accuracy": 0.9978462645882055, + "accuracy": 0.9978547062920896, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -23026,7 +23026,7 @@ } }, { - "accuracy": 0.9976716335666808, + "accuracy": 0.9976800833093492, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -23078,7 +23078,7 @@ } }, { - "accuracy": 0.9979639859183839, + "accuracy": 0.9979718553000375, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -23130,7 +23130,7 @@ } }, { - "accuracy": 0.9988099759150493, + "accuracy": 0.998813913056725, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -23182,7 +23182,7 @@ } }, { - "accuracy": 0.9989841596076363, + "accuracy": 0.9989878253913239, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -23234,7 +23234,7 @@ } }, { - "accuracy": 0.9993571302804508, + "accuracy": 0.9993592321284508, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -23277,7 +23277,7 @@ } }, { - "accuracy": 0.9993914695652691, + "accuracy": 0.9993936185676017, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -23326,7 +23326,7 @@ } }, { - "accuracy": 0.9995273274222487, + "accuracy": 0.9995294323093012, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -23372,7 +23372,7 @@ } }, { - "accuracy": 0.9998241347064706, + "accuracy": 0.9998247221626929, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -23414,7 +23414,7 @@ ], "model.layers.12.self_attn": [ { - "accuracy": 0.9934173626335043, + "accuracy": 0.9934420087619832, "total_bits": 320921600, "q_proj": { "group_size": { @@ -23478,7 +23478,7 @@ } }, { - "accuracy": 0.9939907134363526, + "accuracy": 0.994008483463212, "total_bits": 329244672, "q_proj": { "group_size": { @@ -23542,7 +23542,7 @@ } }, { - "accuracy": 0.9944409986859873, + "accuracy": 0.9944517949694082, "total_bits": 336188416, "q_proj": { "group_size": { @@ -23606,7 +23606,7 @@ } }, { - "accuracy": 0.9963357726994314, + "accuracy": 0.9963408406627806, "total_bits": 401721344, "q_proj": { "group_size": { @@ -23670,7 +23670,7 @@ } }, { - "accuracy": 0.9968768991529942, + "accuracy": 0.9968841179813209, "total_bits": 475443200, "q_proj": { "group_size": { @@ -23734,7 +23734,7 @@ } }, { - "accuracy": 0.9969282485544682, + "accuracy": 0.9969361320530113, "total_bits": 475642880, "q_proj": { "group_size": { @@ -23798,7 +23798,7 @@ } }, { - "accuracy": 0.9979789288420426, + "accuracy": 0.9979799478069732, "total_bits": 609923072, "q_proj": { "group_size": { @@ -23850,7 +23850,7 @@ } }, { - "accuracy": 0.9980530848628596, + "accuracy": 0.9980563888032186, "total_bits": 610188288, "q_proj": { "group_size": { @@ -23902,7 +23902,7 @@ } }, { - "accuracy": 0.9981983447153318, + "accuracy": 0.9982006018491167, "total_bits": 615184384, "q_proj": { "group_size": { @@ -23954,7 +23954,7 @@ } }, { - "accuracy": 0.9982892780711776, + "accuracy": 0.9982906679965948, "total_bits": 624115712, "q_proj": { "group_size": { @@ -24006,7 +24006,7 @@ } }, { - "accuracy": 0.9984227988476816, + "accuracy": 0.9984264558083132, "total_bits": 626637824, "q_proj": { "group_size": { @@ -24070,7 +24070,7 @@ } }, { - "accuracy": 0.9985359768922392, + "accuracy": 0.9985386750807888, "total_bits": 630519808, "q_proj": { "group_size": { @@ -24134,7 +24134,7 @@ } }, { - "accuracy": 0.998591328824037, + "accuracy": 0.9985953500788463, "total_bits": 637526016, "q_proj": { "group_size": { @@ -24195,7 +24195,7 @@ } }, { - "accuracy": 0.9987153888336922, + "accuracy": 0.9987188960964742, "total_bits": 646987776, "q_proj": { "group_size": { @@ -24256,7 +24256,7 @@ } }, { - "accuracy": 0.9992253832322987, + "accuracy": 0.9992272017621681, "total_bits": 784904192, "q_proj": { "group_size": { @@ -24317,7 +24317,7 @@ } }, { - "accuracy": 0.9993441093124842, + "accuracy": 0.9993458338277904, "total_bits": 797982720, "q_proj": { "group_size": { @@ -24378,7 +24378,7 @@ } }, { - "accuracy": 0.999464456874289, + "accuracy": 0.9994654648594166, "total_bits": 911912960, "q_proj": { "group_size": { @@ -24430,7 +24430,7 @@ } }, { - "accuracy": 0.9996392959728837, + "accuracy": 0.9996408589037233, "total_bits": 942882816, "q_proj": { "group_size": { @@ -24482,7 +24482,7 @@ } }, { - "accuracy": 0.9998475441120958, + "accuracy": 0.9998483477699521, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -24536,7 +24536,7 @@ ], "model.layers.12.mlp": [ { - "accuracy": 0.9896305572045477, + "accuracy": 0.9896678955931413, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -24588,7 +24588,7 @@ } }, { - "accuracy": 0.9899074717571861, + "accuracy": 0.9899449050426483, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -24640,7 +24640,7 @@ } }, { - "accuracy": 0.9915675834605568, + "accuracy": 0.9916016043801057, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -24689,7 +24689,7 @@ } }, { - "accuracy": 0.9920901930645892, + "accuracy": 0.9921237355784366, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -24738,7 +24738,7 @@ } }, { - "accuracy": 0.9949220971841561, + "accuracy": 0.9949396244789425, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -24790,7 +24790,7 @@ } }, { - "accuracy": 0.9953178994749722, + "accuracy": 0.9953342527151108, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -24842,7 +24842,7 @@ } }, { - "accuracy": 0.9960481132331648, + "accuracy": 0.9960622744340646, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -24891,7 +24891,7 @@ } }, { - "accuracy": 0.9973745322541187, + "accuracy": 0.997383245119923, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -24934,7 +24934,7 @@ } }, { - "accuracy": 0.9976193295105508, + "accuracy": 0.9976274586821857, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -24977,7 +24977,7 @@ } }, { - "accuracy": 0.9974273035400792, + "accuracy": 0.9974347603948492, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -25029,7 +25029,7 @@ } }, { - "accuracy": 0.9977453800016328, + "accuracy": 0.9977533223205491, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -25081,7 +25081,7 @@ } }, { - "accuracy": 0.9986875814042593, + "accuracy": 0.9986915336431641, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -25133,7 +25133,7 @@ } }, { - "accuracy": 0.9988761284437618, + "accuracy": 0.9988803702749705, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -25185,7 +25185,7 @@ } }, { - "accuracy": 0.9992938578422916, + "accuracy": 0.9992958656757286, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -25228,7 +25228,7 @@ } }, { - "accuracy": 0.9993300135002324, + "accuracy": 0.9993323544530492, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -25277,7 +25277,7 @@ } }, { - "accuracy": 0.9994794491580442, + "accuracy": 0.9994816131595718, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -25323,7 +25323,7 @@ } }, { - "accuracy": 0.9998086616274362, + "accuracy": 0.9998093195088011, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -25365,7 +25365,7 @@ ], "model.layers.13.self_attn": [ { - "accuracy": 0.993070269885816, + "accuracy": 0.9931013097888545, "total_bits": 320921600, "q_proj": { "group_size": { @@ -25429,7 +25429,7 @@ } }, { - "accuracy": 0.9935942409854186, + "accuracy": 0.9936303186573481, "total_bits": 329244672, "q_proj": { "group_size": { @@ -25493,7 +25493,7 @@ } }, { - "accuracy": 0.9942601887803328, + "accuracy": 0.9942820150601236, "total_bits": 336188416, "q_proj": { "group_size": { @@ -25557,7 +25557,7 @@ } }, { - "accuracy": 0.996143207738274, + "accuracy": 0.9961625231724036, "total_bits": 401721344, "q_proj": { "group_size": { @@ -25621,7 +25621,7 @@ } }, { - "accuracy": 0.9966762446259197, + "accuracy": 0.9966877502830405, "total_bits": 475443200, "q_proj": { "group_size": { @@ -25685,7 +25685,7 @@ } }, { - "accuracy": 0.9967599703293097, + "accuracy": 0.9967640918728552, "total_bits": 475642880, "q_proj": { "group_size": { @@ -25749,7 +25749,7 @@ } }, { - "accuracy": 0.9978255539348251, + "accuracy": 0.9978348044188399, "total_bits": 609923072, "q_proj": { "group_size": { @@ -25801,7 +25801,7 @@ } }, { - "accuracy": 0.9979356917503633, + "accuracy": 0.9979336406839522, "total_bits": 610188288, "q_proj": { "group_size": { @@ -25853,7 +25853,7 @@ } }, { - "accuracy": 0.9981025498556463, + "accuracy": 0.9981073982229358, "total_bits": 615184384, "q_proj": { "group_size": { @@ -25905,7 +25905,7 @@ } }, { - "accuracy": 0.9981878578270736, + "accuracy": 0.9981959779795847, "total_bits": 624115712, "q_proj": { "group_size": { @@ -25957,7 +25957,7 @@ } }, { - "accuracy": 0.9983461561955904, + "accuracy": 0.9983563628047705, "total_bits": 626637824, "q_proj": { "group_size": { @@ -26021,7 +26021,7 @@ } }, { - "accuracy": 0.9984465165946045, + "accuracy": 0.9984508518521723, "total_bits": 630519808, "q_proj": { "group_size": { @@ -26085,7 +26085,7 @@ } }, { - "accuracy": 0.9985308884397933, + "accuracy": 0.9985380913866194, "total_bits": 637526016, "q_proj": { "group_size": { @@ -26146,7 +26146,7 @@ } }, { - "accuracy": 0.9986483159622079, + "accuracy": 0.9986517515621687, "total_bits": 646987776, "q_proj": { "group_size": { @@ -26207,7 +26207,7 @@ } }, { - "accuracy": 0.9991856490899074, + "accuracy": 0.9991879906309279, "total_bits": 784904192, "q_proj": { "group_size": { @@ -26268,7 +26268,7 @@ } }, { - "accuracy": 0.999315404774327, + "accuracy": 0.9993169191538503, "total_bits": 797982720, "q_proj": { "group_size": { @@ -26329,7 +26329,7 @@ } }, { - "accuracy": 0.9994201641529799, + "accuracy": 0.9994216143692795, "total_bits": 911912960, "q_proj": { "group_size": { @@ -26381,7 +26381,7 @@ } }, { - "accuracy": 0.9996360661461949, + "accuracy": 0.9996373117901385, "total_bits": 942882816, "q_proj": { "group_size": { @@ -26433,7 +26433,7 @@ } }, { - "accuracy": 0.9998304814860028, + "accuracy": 0.9998308631444448, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -26487,7 +26487,7 @@ ], "model.layers.13.mlp": [ { - "accuracy": 0.988341687541259, + "accuracy": 0.9883852671635779, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -26539,7 +26539,7 @@ } }, { - "accuracy": 0.9886393115708703, + "accuracy": 0.9886814073512429, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -26591,7 +26591,7 @@ } }, { - "accuracy": 0.9904520017536063, + "accuracy": 0.9904850642932089, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -26640,7 +26640,7 @@ } }, { - "accuracy": 0.9910146570519397, + "accuracy": 0.9910452875651812, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -26689,7 +26689,7 @@ } }, { - "accuracy": 0.994352787733078, + "accuracy": 0.9943738771896613, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -26741,7 +26741,7 @@ } }, { - "accuracy": 0.9947901697535264, + "accuracy": 0.994810468272159, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -26793,7 +26793,7 @@ } }, { - "accuracy": 0.9955948802985644, + "accuracy": 0.9956115488159029, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -26842,7 +26842,7 @@ } }, { - "accuracy": 0.9969681954305423, + "accuracy": 0.9969801273393003, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -26885,7 +26885,7 @@ } }, { - "accuracy": 0.9973495753580018, + "accuracy": 0.9973590393599711, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -26928,7 +26928,7 @@ } }, { - "accuracy": 0.9971371177387865, + "accuracy": 0.9971483167456953, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -26980,7 +26980,7 @@ } }, { - "accuracy": 0.9974927717917844, + "accuracy": 0.9975019556127096, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -27032,7 +27032,7 @@ } }, { - "accuracy": 0.9985399655017414, + "accuracy": 0.9985456810773987, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -27084,7 +27084,7 @@ } }, { - "accuracy": 0.9987483353010917, + "accuracy": 0.9987528792729503, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -27136,7 +27136,7 @@ } }, { - "accuracy": 0.9992116867123466, + "accuracy": 0.9992149593308568, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -27179,7 +27179,7 @@ } }, { - "accuracy": 0.9992518643113343, + "accuracy": 0.9992547233736044, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -27228,7 +27228,7 @@ } }, { - "accuracy": 0.9994152595632171, + "accuracy": 0.9994173529802969, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -27274,7 +27274,7 @@ } }, { - "accuracy": 0.9997732858663719, + "accuracy": 0.9997737483170471, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -27316,7 +27316,7 @@ ], "model.layers.14.self_attn": [ { - "accuracy": 0.9913455443946939, + "accuracy": 0.991377169364377, "total_bits": 320921600, "q_proj": { "group_size": { @@ -27380,7 +27380,7 @@ } }, { - "accuracy": 0.992038853074375, + "accuracy": 0.9920557807934912, "total_bits": 329244672, "q_proj": { "group_size": { @@ -27444,7 +27444,7 @@ } }, { - "accuracy": 0.9927284756773397, + "accuracy": 0.992753801377196, "total_bits": 336188416, "q_proj": { "group_size": { @@ -27508,7 +27508,7 @@ } }, { - "accuracy": 0.9953102092993887, + "accuracy": 0.9953258970850393, "total_bits": 401721344, "q_proj": { "group_size": { @@ -27572,7 +27572,7 @@ } }, { - "accuracy": 0.9957935492459097, + "accuracy": 0.9958077418176752, "total_bits": 475443200, "q_proj": { "group_size": { @@ -27636,7 +27636,7 @@ } }, { - "accuracy": 0.9958985463569039, + "accuracy": 0.9959129593090007, "total_bits": 475642880, "q_proj": { "group_size": { @@ -27700,7 +27700,7 @@ } }, { - "accuracy": 0.997289663082675, + "accuracy": 0.9973009240470434, "total_bits": 609923072, "q_proj": { "group_size": { @@ -27752,7 +27752,7 @@ } }, { - "accuracy": 0.997442047062673, + "accuracy": 0.9974573403596878, "total_bits": 610188288, "q_proj": { "group_size": { @@ -27804,7 +27804,7 @@ } }, { - "accuracy": 0.9976123914514717, + "accuracy": 0.9976206012070179, "total_bits": 615184384, "q_proj": { "group_size": { @@ -27856,7 +27856,7 @@ } }, { - "accuracy": 0.9977303186529561, + "accuracy": 0.9977403832109351, "total_bits": 624115712, "q_proj": { "group_size": { @@ -27908,7 +27908,7 @@ } }, { - "accuracy": 0.9979208405864867, + "accuracy": 0.9979291446506977, "total_bits": 626637824, "q_proj": { "group_size": { @@ -27972,7 +27972,7 @@ } }, { - "accuracy": 0.9980615659763938, + "accuracy": 0.9980669735293639, "total_bits": 630519808, "q_proj": { "group_size": { @@ -28036,7 +28036,7 @@ } }, { - "accuracy": 0.9981739960218731, + "accuracy": 0.9981791310404476, "total_bits": 637526016, "q_proj": { "group_size": { @@ -28097,7 +28097,7 @@ } }, { - "accuracy": 0.9983291598527055, + "accuracy": 0.9983346197558077, "total_bits": 646987776, "q_proj": { "group_size": { @@ -28158,7 +28158,7 @@ } }, { - "accuracy": 0.9989834496457326, + "accuracy": 0.9989880877301881, "total_bits": 784904192, "q_proj": { "group_size": { @@ -28219,7 +28219,7 @@ } }, { - "accuracy": 0.999152531357188, + "accuracy": 0.9991554396325036, "total_bits": 797982720, "q_proj": { "group_size": { @@ -28280,7 +28280,7 @@ } }, { - "accuracy": 0.9992825815356091, + "accuracy": 0.9992862485937382, "total_bits": 911912960, "q_proj": { "group_size": { @@ -28332,7 +28332,7 @@ } }, { - "accuracy": 0.9995438568784218, + "accuracy": 0.9995451565635832, "total_bits": 942882816, "q_proj": { "group_size": { @@ -28384,7 +28384,7 @@ } }, { - "accuracy": 0.9998026666304979, + "accuracy": 0.9998035634941372, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -28438,7 +28438,7 @@ ], "model.layers.14.mlp": [ { - "accuracy": 0.987615726496044, + "accuracy": 0.987659760995915, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -28490,7 +28490,7 @@ } }, { - "accuracy": 0.9879262706166819, + "accuracy": 0.9879651132382845, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -28542,7 +28542,7 @@ } }, { - "accuracy": 0.9899385893031171, + "accuracy": 0.9899712542169973, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -28591,7 +28591,7 @@ } }, { - "accuracy": 0.9905699828737661, + "accuracy": 0.9906006596590343, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -28640,7 +28640,7 @@ } }, { - "accuracy": 0.9939034424330059, + "accuracy": 0.9939244464039803, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -28692,7 +28692,7 @@ } }, { - "accuracy": 0.9943796988380583, + "accuracy": 0.9943985405721163, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -28744,7 +28744,7 @@ } }, { - "accuracy": 0.9952624318631071, + "accuracy": 0.9952786890299696, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -28793,7 +28793,7 @@ } }, { - "accuracy": 0.9968507274201042, + "accuracy": 0.9968623528747182, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -28836,7 +28836,7 @@ } }, { - "accuracy": 0.9971463515570289, + "accuracy": 0.9971565402259952, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -28879,7 +28879,7 @@ } }, { - "accuracy": 0.9969085562778147, + "accuracy": 0.9969193956961757, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -28931,7 +28931,7 @@ } }, { - "accuracy": 0.9972926278255487, + "accuracy": 0.9973022090761285, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -28983,7 +28983,7 @@ } }, { - "accuracy": 0.9984239709416503, + "accuracy": 0.9984294921159744, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -29035,7 +29035,7 @@ } }, { - "accuracy": 0.9986506122115412, + "accuracy": 0.9986554425405828, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -29087,7 +29087,7 @@ } }, { - "accuracy": 0.9991530533879995, + "accuracy": 0.9991561776321185, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -29130,7 +29130,7 @@ } }, { - "accuracy": 0.9991957375191545, + "accuracy": 0.9991987862771279, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -29179,7 +29179,7 @@ } }, { - "accuracy": 0.9993766586443311, + "accuracy": 0.9993786867707968, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -29225,7 +29225,7 @@ } }, { - "accuracy": 0.9997690551130003, + "accuracy": 0.9997700279774634, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -29267,7 +29267,7 @@ ], "model.layers.15.self_attn": [ { - "accuracy": 0.9928274546798906, + "accuracy": 0.9928672023509678, "total_bits": 320921600, "q_proj": { "group_size": { @@ -29331,7 +29331,7 @@ } }, { - "accuracy": 0.9933157625951266, + "accuracy": 0.9933554749739798, "total_bits": 329244672, "q_proj": { "group_size": { @@ -29395,7 +29395,7 @@ } }, { - "accuracy": 0.9940276510621372, + "accuracy": 0.9940509039320444, "total_bits": 336188416, "q_proj": { "group_size": { @@ -29459,7 +29459,7 @@ } }, { - "accuracy": 0.9957418775087908, + "accuracy": 0.9957557936248026, "total_bits": 401721344, "q_proj": { "group_size": { @@ -29523,7 +29523,7 @@ } }, { - "accuracy": 0.9965126796772605, + "accuracy": 0.9965165794679993, "total_bits": 475443200, "q_proj": { "group_size": { @@ -29587,7 +29587,7 @@ } }, { - "accuracy": 0.9966262116243965, + "accuracy": 0.9966490264785918, "total_bits": 475642880, "q_proj": { "group_size": { @@ -29651,7 +29651,7 @@ } }, { - "accuracy": 0.9975820511187378, + "accuracy": 0.9975831069444355, "total_bits": 609923072, "q_proj": { "group_size": { @@ -29703,7 +29703,7 @@ } }, { - "accuracy": 0.9977370616244642, + "accuracy": 0.9977637364676124, "total_bits": 610188288, "q_proj": { "group_size": { @@ -29755,7 +29755,7 @@ } }, { - "accuracy": 0.9980304082365412, + "accuracy": 0.9980394099103777, "total_bits": 615184384, "q_proj": { "group_size": { @@ -29807,7 +29807,7 @@ } }, { - "accuracy": 0.9981127049969999, + "accuracy": 0.9981207359386118, "total_bits": 624115712, "q_proj": { "group_size": { @@ -29859,7 +29859,7 @@ } }, { - "accuracy": 0.9982812486981091, + "accuracy": 0.9982769618693151, "total_bits": 626637824, "q_proj": { "group_size": { @@ -29923,7 +29923,7 @@ } }, { - "accuracy": 0.9984158369663515, + "accuracy": 0.9984156839353474, "total_bits": 630519808, "q_proj": { "group_size": { @@ -29987,7 +29987,7 @@ } }, { - "accuracy": 0.9984464180704794, + "accuracy": 0.9984421201638485, "total_bits": 637526016, "q_proj": { "group_size": { @@ -30048,7 +30048,7 @@ } }, { - "accuracy": 0.9985908895338836, + "accuracy": 0.9985888167039344, "total_bits": 646987776, "q_proj": { "group_size": { @@ -30109,7 +30109,7 @@ } }, { - "accuracy": 0.9991415730236393, + "accuracy": 0.9991463512966507, "total_bits": 784904192, "q_proj": { "group_size": { @@ -30170,7 +30170,7 @@ } }, { - "accuracy": 0.9992970812477564, + "accuracy": 0.9992999401629755, "total_bits": 797982720, "q_proj": { "group_size": { @@ -30231,7 +30231,7 @@ } }, { - "accuracy": 0.9993664823276432, + "accuracy": 0.9993709608618366, "total_bits": 911912960, "q_proj": { "group_size": { @@ -30283,7 +30283,7 @@ } }, { - "accuracy": 0.9996263132381596, + "accuracy": 0.9996272085822726, "total_bits": 942882816, "q_proj": { "group_size": { @@ -30335,7 +30335,7 @@ } }, { - "accuracy": 0.9998181497947755, + "accuracy": 0.9998190558368438, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -30389,7 +30389,7 @@ ], "model.layers.15.mlp": [ { - "accuracy": 0.9861827853478884, + "accuracy": 0.9862336375211415, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -30441,7 +30441,7 @@ } }, { - "accuracy": 0.9865610176011136, + "accuracy": 0.9866093287342473, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -30493,7 +30493,7 @@ } }, { - "accuracy": 0.9887580455918061, + "accuracy": 0.9887923355165281, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -30542,7 +30542,7 @@ } }, { - "accuracy": 0.9894449969655589, + "accuracy": 0.9894778650057944, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -30591,7 +30591,7 @@ } }, { - "accuracy": 0.9931852331286982, + "accuracy": 0.993210820775283, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -30643,7 +30643,7 @@ } }, { - "accuracy": 0.9937151529287037, + "accuracy": 0.9937371690022317, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -30695,7 +30695,7 @@ } }, { - "accuracy": 0.9946746175226412, + "accuracy": 0.9946933933778813, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -30744,7 +30744,7 @@ } }, { - "accuracy": 0.9964694561142671, + "accuracy": 0.996483312625634, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -30787,7 +30787,7 @@ } }, { - "accuracy": 0.9967968228616213, + "accuracy": 0.9968091486708114, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -30830,7 +30830,7 @@ } }, { - "accuracy": 0.9965471625328064, + "accuracy": 0.996560501032754, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -30882,7 +30882,7 @@ } }, { - "accuracy": 0.9969742698104758, + "accuracy": 0.9969857056674204, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -30934,7 +30934,7 @@ } }, { - "accuracy": 0.9982423123560453, + "accuracy": 0.9982490100358662, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -30986,7 +30986,7 @@ } }, { - "accuracy": 0.9984931288973281, + "accuracy": 0.9984991175955847, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -31038,7 +31038,7 @@ } }, { - "accuracy": 0.999052910820434, + "accuracy": 0.9990566644425455, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -31081,7 +31081,7 @@ } }, { - "accuracy": 0.9991022469966035, + "accuracy": 0.9991061164948501, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -31130,7 +31130,7 @@ } }, { - "accuracy": 0.9992966252428136, + "accuracy": 0.9992995234696489, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -31176,7 +31176,7 @@ } }, { - "accuracy": 0.9997429649198526, + "accuracy": 0.9997439644005346, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -31218,7 +31218,7 @@ ], "model.layers.16.self_attn": [ { - "accuracy": 0.9925968192125622, + "accuracy": 0.9925946985420427, "total_bits": 320921600, "q_proj": { "group_size": { @@ -31282,7 +31282,7 @@ } }, { - "accuracy": 0.99309735705978, + "accuracy": 0.9930919691136009, "total_bits": 329244672, "q_proj": { "group_size": { @@ -31346,7 +31346,7 @@ } }, { - "accuracy": 0.9937743744567821, + "accuracy": 0.9937621939339136, "total_bits": 336188416, "q_proj": { "group_size": { @@ -31410,7 +31410,7 @@ } }, { - "accuracy": 0.9956669995659276, + "accuracy": 0.9956522540826547, "total_bits": 401721344, "q_proj": { "group_size": { @@ -31474,7 +31474,7 @@ } }, { - "accuracy": 0.9963800652246726, + "accuracy": 0.9963778065223443, "total_bits": 475443200, "q_proj": { "group_size": { @@ -31538,7 +31538,7 @@ } }, { - "accuracy": 0.996505618095398, + "accuracy": 0.9965122761695009, "total_bits": 475642880, "q_proj": { "group_size": { @@ -31602,7 +31602,7 @@ } }, { - "accuracy": 0.9975578698672747, + "accuracy": 0.9975628733242813, "total_bits": 609923072, "q_proj": { "group_size": { @@ -31654,7 +31654,7 @@ } }, { - "accuracy": 0.997743582058894, + "accuracy": 0.9977623024269154, "total_bits": 610188288, "q_proj": { "group_size": { @@ -31706,7 +31706,7 @@ } }, { - "accuracy": 0.9980485251075343, + "accuracy": 0.9980577240256887, "total_bits": 615184384, "q_proj": { "group_size": { @@ -31758,7 +31758,7 @@ } }, { - "accuracy": 0.9981357082724571, + "accuracy": 0.9981441109588272, "total_bits": 624115712, "q_proj": { "group_size": { @@ -31810,7 +31810,7 @@ } }, { - "accuracy": 0.9982351198008186, + "accuracy": 0.9982418553217461, "total_bits": 626637824, "q_proj": { "group_size": { @@ -31874,7 +31874,7 @@ } }, { - "accuracy": 0.9983807114982292, + "accuracy": 0.9983799330105907, "total_bits": 630519808, "q_proj": { "group_size": { @@ -31938,7 +31938,7 @@ } }, { - "accuracy": 0.9984220210462809, + "accuracy": 0.9984260246549782, "total_bits": 637526016, "q_proj": { "group_size": { @@ -31999,7 +31999,7 @@ } }, { - "accuracy": 0.9985868825920319, + "accuracy": 0.9985859461716915, "total_bits": 646987776, "q_proj": { "group_size": { @@ -32060,7 +32060,7 @@ } }, { - "accuracy": 0.9991260623853457, + "accuracy": 0.9991273264351644, "total_bits": 784904192, "q_proj": { "group_size": { @@ -32121,7 +32121,7 @@ } }, { - "accuracy": 0.9992886570919501, + "accuracy": 0.9992887959080307, "total_bits": 797982720, "q_proj": { "group_size": { @@ -32182,7 +32182,7 @@ } }, { - "accuracy": 0.9993726274880924, + "accuracy": 0.9993739917286133, "total_bits": 911912960, "q_proj": { "group_size": { @@ -32234,7 +32234,7 @@ } }, { - "accuracy": 0.9996211085419514, + "accuracy": 0.9996216888294408, "total_bits": 942882816, "q_proj": { "group_size": { @@ -32286,7 +32286,7 @@ } }, { - "accuracy": 0.9998230821158933, + "accuracy": 0.9998234613725033, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -32340,7 +32340,7 @@ ], "model.layers.16.mlp": [ { - "accuracy": 0.9849794836420762, + "accuracy": 0.9850243957419145, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -32392,7 +32392,7 @@ } }, { - "accuracy": 0.9853776709029549, + "accuracy": 0.9854216716791454, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -32444,7 +32444,7 @@ } }, { - "accuracy": 0.9877158207328696, + "accuracy": 0.9877575651595467, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -32493,7 +32493,7 @@ } }, { - "accuracy": 0.988448208884189, + "accuracy": 0.9884875142260602, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -32542,7 +32542,7 @@ } }, { - "accuracy": 0.9926008731126785, + "accuracy": 0.9926245706646066, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -32594,7 +32594,7 @@ } }, { - "accuracy": 0.9931793134463461, + "accuracy": 0.9932013044231817, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -32646,7 +32646,7 @@ } }, { - "accuracy": 0.9942079453091872, + "accuracy": 0.9942263188330751, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -32695,7 +32695,7 @@ } }, { - "accuracy": 0.9961664115912036, + "accuracy": 0.9961791775728527, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -32738,7 +32738,7 @@ } }, { - "accuracy": 0.9965244966902231, + "accuracy": 0.9965349733829498, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -32781,7 +32781,7 @@ } }, { - "accuracy": 0.9962456234191593, + "accuracy": 0.9962578137454233, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -32833,7 +32833,7 @@ } }, { - "accuracy": 0.9967111456943186, + "accuracy": 0.9967214327893759, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -32885,7 +32885,7 @@ } }, { - "accuracy": 0.9980846478750831, + "accuracy": 0.9980908447974607, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -32937,7 +32937,7 @@ } }, { - "accuracy": 0.9983606576723488, + "accuracy": 0.9983657353410595, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -32989,7 +32989,7 @@ } }, { - "accuracy": 0.9989680404725828, + "accuracy": 0.9989717606651155, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -33032,7 +33032,7 @@ } }, { - "accuracy": 0.9990227958677631, + "accuracy": 0.9990256862028649, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -33081,7 +33081,7 @@ } }, { - "accuracy": 0.9992343351539028, + "accuracy": 0.9992367130655208, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -33127,7 +33127,7 @@ } }, { - "accuracy": 0.9997213545598482, + "accuracy": 0.999722396905877, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -33169,7 +33169,7 @@ ], "model.layers.17.self_attn": [ { - "accuracy": 0.9910560665946258, + "accuracy": 0.9910892216782821, "total_bits": 320921600, "q_proj": { "group_size": { @@ -33233,7 +33233,7 @@ } }, { - "accuracy": 0.9917917706464466, + "accuracy": 0.9918120174031508, "total_bits": 329244672, "q_proj": { "group_size": { @@ -33297,7 +33297,7 @@ } }, { - "accuracy": 0.9926533079461047, + "accuracy": 0.9926765067012686, "total_bits": 336188416, "q_proj": { "group_size": { @@ -33361,7 +33361,7 @@ } }, { - "accuracy": 0.9949257083629307, + "accuracy": 0.9949391390147962, "total_bits": 401721344, "q_proj": { "group_size": { @@ -33425,7 +33425,7 @@ } }, { - "accuracy": 0.9955483655396261, + "accuracy": 0.9955674139292616, "total_bits": 475443200, "q_proj": { "group_size": { @@ -33489,7 +33489,7 @@ } }, { - "accuracy": 0.9957950672036723, + "accuracy": 0.9958150167214243, "total_bits": 475642880, "q_proj": { "group_size": { @@ -33553,7 +33553,7 @@ } }, { - "accuracy": 0.9968679990423354, + "accuracy": 0.9968880228698254, "total_bits": 609923072, "q_proj": { "group_size": { @@ -33605,7 +33605,7 @@ } }, { - "accuracy": 0.9972093024928319, + "accuracy": 0.9972249775340682, "total_bits": 610188288, "q_proj": { "group_size": { @@ -33657,7 +33657,7 @@ } }, { - "accuracy": 0.9975203907019213, + "accuracy": 0.9975321314444667, "total_bits": 615184384, "q_proj": { "group_size": { @@ -33709,7 +33709,7 @@ } }, { - "accuracy": 0.9976376671540109, + "accuracy": 0.9976504441154631, "total_bits": 624115712, "q_proj": { "group_size": { @@ -33761,7 +33761,7 @@ } }, { - "accuracy": 0.9978677502980358, + "accuracy": 0.9978752383276036, "total_bits": 626637824, "q_proj": { "group_size": { @@ -33825,7 +33825,7 @@ } }, { - "accuracy": 0.9980360889121106, + "accuracy": 0.9980434369491903, "total_bits": 630519808, "q_proj": { "group_size": { @@ -33889,7 +33889,7 @@ } }, { - "accuracy": 0.9981242753565311, + "accuracy": 0.9981333668294706, "total_bits": 637526016, "q_proj": { "group_size": { @@ -33950,7 +33950,7 @@ } }, { - "accuracy": 0.998312907195405, + "accuracy": 0.9983174836165026, "total_bits": 646987776, "q_proj": { "group_size": { @@ -34011,7 +34011,7 @@ } }, { - "accuracy": 0.9989461348833222, + "accuracy": 0.9989511106947535, "total_bits": 784904192, "q_proj": { "group_size": { @@ -34072,7 +34072,7 @@ } }, { - "accuracy": 0.9991515556252316, + "accuracy": 0.9991542639112786, "total_bits": 797982720, "q_proj": { "group_size": { @@ -34133,7 +34133,7 @@ } }, { - "accuracy": 0.9992028865963221, + "accuracy": 0.9992061690182278, "total_bits": 911912960, "q_proj": { "group_size": { @@ -34185,7 +34185,7 @@ } }, { - "accuracy": 0.9995281457116729, + "accuracy": 0.9995295110795843, "total_bits": 942882816, "q_proj": { "group_size": { @@ -34237,7 +34237,7 @@ } }, { - "accuracy": 0.9997763727350455, + "accuracy": 0.9997765623817318, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -34291,7 +34291,7 @@ ], "model.layers.17.mlp": [ { - "accuracy": 0.984804674198753, + "accuracy": 0.9848562996638449, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -34343,7 +34343,7 @@ } }, { - "accuracy": 0.9851924714289213, + "accuracy": 0.9852418428973148, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -34395,7 +34395,7 @@ } }, { - "accuracy": 0.987509338479293, + "accuracy": 0.9875535047367999, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -34444,7 +34444,7 @@ } }, { - "accuracy": 0.98826927417203, + "accuracy": 0.9883107770430414, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -34493,7 +34493,7 @@ } }, { - "accuracy": 0.9924520552158356, + "accuracy": 0.9924786514357516, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -34545,7 +34545,7 @@ } }, { - "accuracy": 0.9930413851612493, + "accuracy": 0.9930654787703564, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -34597,7 +34597,7 @@ } }, { - "accuracy": 0.994079129868432, + "accuracy": 0.9940989880185378, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -34646,7 +34646,7 @@ } }, { - "accuracy": 0.9960772140245688, + "accuracy": 0.9960923030188209, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -34689,7 +34689,7 @@ } }, { - "accuracy": 0.9964417611297808, + "accuracy": 0.9964542581062568, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -34732,7 +34732,7 @@ } }, { - "accuracy": 0.9961744750800886, + "accuracy": 0.9961885747156645, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -34784,7 +34784,7 @@ } }, { - "accuracy": 0.9966468340472171, + "accuracy": 0.9966585283216677, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -34836,7 +34836,7 @@ } }, { - "accuracy": 0.99805246646467, + "accuracy": 0.9980594692261595, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -34888,7 +34888,7 @@ } }, { - "accuracy": 0.998329305335095, + "accuracy": 0.9983351018867994, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -34940,7 +34940,7 @@ } }, { - "accuracy": 0.9989466951473763, + "accuracy": 0.9989502415257064, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -34983,7 +34983,7 @@ } }, { - "accuracy": 0.9990089835697099, + "accuracy": 0.9990120304650382, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -35032,7 +35032,7 @@ } }, { - "accuracy": 0.9992218252369448, + "accuracy": 0.9992244545076239, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -35078,7 +35078,7 @@ } }, { - "accuracy": 0.9997136498507309, + "accuracy": 0.9997146602131819, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -35120,7 +35120,7 @@ ], "model.layers.18.self_attn": [ { - "accuracy": 0.9913529447819057, + "accuracy": 0.9913982756828007, "total_bits": 320921600, "q_proj": { "group_size": { @@ -35184,7 +35184,7 @@ } }, { - "accuracy": 0.9920830067835356, + "accuracy": 0.9921223301636545, "total_bits": 329244672, "q_proj": { "group_size": { @@ -35248,7 +35248,7 @@ } }, { - "accuracy": 0.9929305134635222, + "accuracy": 0.9929613604357368, "total_bits": 336188416, "q_proj": { "group_size": { @@ -35312,7 +35312,7 @@ } }, { - "accuracy": 0.9946415000840237, + "accuracy": 0.9946649003969995, "total_bits": 401721344, "q_proj": { "group_size": { @@ -35376,7 +35376,7 @@ } }, { - "accuracy": 0.995871060772946, + "accuracy": 0.9958898327068278, "total_bits": 475443200, "q_proj": { "group_size": { @@ -35440,7 +35440,7 @@ } }, { - "accuracy": 0.9960022133431936, + "accuracy": 0.9960212660463232, "total_bits": 475642880, "q_proj": { "group_size": { @@ -35504,7 +35504,7 @@ } }, { - "accuracy": 0.9968859521966231, + "accuracy": 0.9969053470382565, "total_bits": 609923072, "q_proj": { "group_size": { @@ -35556,7 +35556,7 @@ } }, { - "accuracy": 0.997050899620119, + "accuracy": 0.9970676089195829, "total_bits": 610188288, "q_proj": { "group_size": { @@ -35608,7 +35608,7 @@ } }, { - "accuracy": 0.9972855605577168, + "accuracy": 0.9973031739263158, "total_bits": 615184384, "q_proj": { "group_size": { @@ -35660,7 +35660,7 @@ } }, { - "accuracy": 0.9974472483522013, + "accuracy": 0.9974496664577409, "total_bits": 624115712, "q_proj": { "group_size": { @@ -35712,7 +35712,7 @@ } }, { - "accuracy": 0.9979674829856345, + "accuracy": 0.997974093807371, "total_bits": 626637824, "q_proj": { "group_size": { @@ -35776,7 +35776,7 @@ } }, { - "accuracy": 0.9981001360636008, + "accuracy": 0.9981133094743678, "total_bits": 630519808, "q_proj": { "group_size": { @@ -35840,7 +35840,7 @@ } }, { - "accuracy": 0.9981795843494566, + "accuracy": 0.9981874741221729, "total_bits": 637526016, "q_proj": { "group_size": { @@ -35901,7 +35901,7 @@ } }, { - "accuracy": 0.9983350740451562, + "accuracy": 0.9983470728130717, "total_bits": 646987776, "q_proj": { "group_size": { @@ -35962,7 +35962,7 @@ } }, { - "accuracy": 0.9990033815173727, + "accuracy": 0.9990082317473072, "total_bits": 784904192, "q_proj": { "group_size": { @@ -36023,7 +36023,7 @@ } }, { - "accuracy": 0.9991648680481472, + "accuracy": 0.9991680110167516, "total_bits": 797982720, "q_proj": { "group_size": { @@ -36084,7 +36084,7 @@ } }, { - "accuracy": 0.9991980236709902, + "accuracy": 0.9992027371435573, "total_bits": 911912960, "q_proj": { "group_size": { @@ -36136,7 +36136,7 @@ } }, { - "accuracy": 0.99946588209193, + "accuracy": 0.9994677605205461, "total_bits": 942882816, "q_proj": { "group_size": { @@ -36188,7 +36188,7 @@ } }, { - "accuracy": 0.9997663760165635, + "accuracy": 0.999767539728629, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -36242,7 +36242,7 @@ ], "model.layers.18.mlp": [ { - "accuracy": 0.9828069445333982, + "accuracy": 0.9828674479534751, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -36294,7 +36294,7 @@ } }, { - "accuracy": 0.983248095763357, + "accuracy": 0.9833091513106698, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -36346,7 +36346,7 @@ } }, { - "accuracy": 0.9857654759758397, + "accuracy": 0.9858214149349614, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -36395,7 +36395,7 @@ } }, { - "accuracy": 0.9865966847068385, + "accuracy": 0.9866502081093035, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -36444,7 +36444,7 @@ } }, { - "accuracy": 0.9914528264811164, + "accuracy": 0.9914823582297877, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -36496,7 +36496,7 @@ } }, { - "accuracy": 0.9921184699786337, + "accuracy": 0.9921456269527736, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -36548,7 +36548,7 @@ } }, { - "accuracy": 0.9932489410827035, + "accuracy": 0.993272383746348, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -36597,7 +36597,7 @@ } }, { - "accuracy": 0.9955660363561228, + "accuracy": 0.9955827268330675, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -36640,7 +36640,7 @@ } }, { - "accuracy": 0.9959747179558402, + "accuracy": 0.9959893108982789, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -36683,7 +36683,7 @@ } }, { - "accuracy": 0.9956668242812157, + "accuracy": 0.9956827900911632, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -36735,7 +36735,7 @@ } }, { - "accuracy": 0.9962029719823285, + "accuracy": 0.9962154081777522, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -36787,7 +36787,7 @@ } }, { - "accuracy": 0.9977937921097404, + "accuracy": 0.9978011748508403, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -36839,7 +36839,7 @@ } }, { - "accuracy": 0.9981081668091448, + "accuracy": 0.9981141621737104, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -36891,7 +36891,7 @@ } }, { - "accuracy": 0.9988083807065299, + "accuracy": 0.9988127598833096, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -36934,7 +36934,7 @@ } }, { - "accuracy": 0.998875937669685, + "accuracy": 0.9988800256856178, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -36983,7 +36983,7 @@ } }, { - "accuracy": 0.999104227478567, + "accuracy": 0.9991076134733463, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -37029,7 +37029,7 @@ } }, { - "accuracy": 0.999676902190243, + "accuracy": 0.9996782013852346, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -37071,7 +37071,7 @@ ], "model.layers.19.self_attn": [ { - "accuracy": 0.990196442133502, + "accuracy": 0.990233199376809, "total_bits": 320921600, "q_proj": { "group_size": { @@ -37135,7 +37135,7 @@ } }, { - "accuracy": 0.990905330369347, + "accuracy": 0.9909486911798778, "total_bits": 329244672, "q_proj": { "group_size": { @@ -37199,7 +37199,7 @@ } }, { - "accuracy": 0.9918413632794431, + "accuracy": 0.9918860040212932, "total_bits": 336188416, "q_proj": { "group_size": { @@ -37263,7 +37263,7 @@ } }, { - "accuracy": 0.9941160416132525, + "accuracy": 0.9941537744904819, "total_bits": 401721344, "q_proj": { "group_size": { @@ -37327,7 +37327,7 @@ } }, { - "accuracy": 0.9953031504624769, + "accuracy": 0.9953181806363558, "total_bits": 475443200, "q_proj": { "group_size": { @@ -37391,7 +37391,7 @@ } }, { - "accuracy": 0.9954040379900682, + "accuracy": 0.9954430586413333, "total_bits": 475642880, "q_proj": { "group_size": { @@ -37455,7 +37455,7 @@ } }, { - "accuracy": 0.9967321498613608, + "accuracy": 0.9967416707230242, "total_bits": 609923072, "q_proj": { "group_size": { @@ -37507,7 +37507,7 @@ } }, { - "accuracy": 0.9968605790483324, + "accuracy": 0.9969020932912827, "total_bits": 610188288, "q_proj": { "group_size": { @@ -37559,7 +37559,7 @@ } }, { - "accuracy": 0.9970996750420645, + "accuracy": 0.9971019835456422, "total_bits": 615184384, "q_proj": { "group_size": { @@ -37611,7 +37611,7 @@ } }, { - "accuracy": 0.9972810439373317, + "accuracy": 0.9972712389732662, "total_bits": 624115712, "q_proj": { "group_size": { @@ -37663,7 +37663,7 @@ } }, { - "accuracy": 0.9976433162626467, + "accuracy": 0.9976704179456359, "total_bits": 626637824, "q_proj": { "group_size": { @@ -37727,7 +37727,7 @@ } }, { - "accuracy": 0.9977995131752992, + "accuracy": 0.9978218337422923, "total_bits": 630519808, "q_proj": { "group_size": { @@ -37791,7 +37791,7 @@ } }, { - "accuracy": 0.9978907351431093, + "accuracy": 0.9979183068008799, "total_bits": 637526016, "q_proj": { "group_size": { @@ -37852,7 +37852,7 @@ } }, { - "accuracy": 0.9980636109646998, + "accuracy": 0.9980808229822862, "total_bits": 646987776, "q_proj": { "group_size": { @@ -37913,7 +37913,7 @@ } }, { - "accuracy": 0.9988304565807706, + "accuracy": 0.9988354022957777, "total_bits": 784904192, "q_proj": { "group_size": { @@ -37974,7 +37974,7 @@ } }, { - "accuracy": 0.9990154104797464, + "accuracy": 0.9990198171060336, "total_bits": 797982720, "q_proj": { "group_size": { @@ -38035,7 +38035,7 @@ } }, { - "accuracy": 0.9991160124344262, + "accuracy": 0.9991181930036921, "total_bits": 911912960, "q_proj": { "group_size": { @@ -38087,7 +38087,7 @@ } }, { - "accuracy": 0.9994320649849741, + "accuracy": 0.9994329822396761, "total_bits": 942882816, "q_proj": { "group_size": { @@ -38139,7 +38139,7 @@ } }, { - "accuracy": 0.9997431296168974, + "accuracy": 0.9997449827282444, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -38193,7 +38193,7 @@ ], "model.layers.19.mlp": [ { - "accuracy": 0.9796082471546373, + "accuracy": 0.9796796895955738, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -38245,7 +38245,7 @@ } }, { - "accuracy": 0.9801257629143564, + "accuracy": 0.9802035846208271, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -38297,7 +38297,7 @@ } }, { - "accuracy": 0.9833829449979883, + "accuracy": 0.9834462843443218, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -38346,7 +38346,7 @@ } }, { - "accuracy": 0.9844279697066859, + "accuracy": 0.9844895491474553, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -38395,7 +38395,7 @@ } }, { - "accuracy": 0.9898923376673147, + "accuracy": 0.9899279420313082, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -38447,7 +38447,7 @@ } }, { - "accuracy": 0.9906982795188302, + "accuracy": 0.9907329215815193, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -38499,7 +38499,7 @@ } }, { - "accuracy": 0.9921448952273318, + "accuracy": 0.9921741556180151, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -38548,7 +38548,7 @@ } }, { - "accuracy": 0.9947183873308333, + "accuracy": 0.9947378890294778, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -38591,7 +38591,7 @@ } }, { - "accuracy": 0.9952618448357833, + "accuracy": 0.995278559233013, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -38634,7 +38634,7 @@ } }, { - "accuracy": 0.994880302564094, + "accuracy": 0.9948962287683236, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -38686,7 +38686,7 @@ } }, { - "accuracy": 0.9954716237752061, + "accuracy": 0.9954882856262358, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -38738,7 +38738,7 @@ } }, { - "accuracy": 0.9973968157642766, + "accuracy": 0.9974060462493646, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -38790,7 +38790,7 @@ } }, { - "accuracy": 0.997750143667585, + "accuracy": 0.9977669859010923, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -38842,7 +38842,7 @@ } }, { - "accuracy": 0.9985973960474918, + "accuracy": 0.9986026950768734, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -38885,7 +38885,7 @@ } }, { - "accuracy": 0.9986736983256904, + "accuracy": 0.9986789017719658, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -38934,7 +38934,7 @@ } }, { - "accuracy": 0.9989733811663953, + "accuracy": 0.9989772184114707, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -38980,7 +38980,7 @@ } }, { - "accuracy": 0.9996190920325094, + "accuracy": 0.9996207495170989, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -39022,7 +39022,7 @@ ], "model.layers.20.self_attn": [ { - "accuracy": 0.9895887900339929, + "accuracy": 0.9896177124035986, "total_bits": 320921600, "q_proj": { "group_size": { @@ -39086,7 +39086,7 @@ } }, { - "accuracy": 0.9903233819886258, + "accuracy": 0.9903603447111029, "total_bits": 329244672, "q_proj": { "group_size": { @@ -39150,7 +39150,7 @@ } }, { - "accuracy": 0.9913752549572995, + "accuracy": 0.991380344095983, "total_bits": 336188416, "q_proj": { "group_size": { @@ -39214,7 +39214,7 @@ } }, { - "accuracy": 0.9943125498922247, + "accuracy": 0.994308623828386, "total_bits": 401721344, "q_proj": { "group_size": { @@ -39278,7 +39278,7 @@ } }, { - "accuracy": 0.9949365226845992, + "accuracy": 0.9949346502360544, "total_bits": 475443200, "q_proj": { "group_size": { @@ -39342,7 +39342,7 @@ } }, { - "accuracy": 0.9950673654675484, + "accuracy": 0.9950743776402975, "total_bits": 475642880, "q_proj": { "group_size": { @@ -39406,7 +39406,7 @@ } }, { - "accuracy": 0.9966826113431078, + "accuracy": 0.9966742263028496, "total_bits": 609923072, "q_proj": { "group_size": { @@ -39458,7 +39458,7 @@ } }, { - "accuracy": 0.996860275731275, + "accuracy": 0.9968655807407278, "total_bits": 610188288, "q_proj": { "group_size": { @@ -39510,7 +39510,7 @@ } }, { - "accuracy": 0.9971656989502279, + "accuracy": 0.9971730654176912, "total_bits": 615184384, "q_proj": { "group_size": { @@ -39562,7 +39562,7 @@ } }, { - "accuracy": 0.997288644902016, + "accuracy": 0.9972955962937129, "total_bits": 624115712, "q_proj": { "group_size": { @@ -39614,7 +39614,7 @@ } }, { - "accuracy": 0.9974949987311112, + "accuracy": 0.9974989298927156, "total_bits": 626637824, "q_proj": { "group_size": { @@ -39678,7 +39678,7 @@ } }, { - "accuracy": 0.9976547374144981, + "accuracy": 0.9976605945511868, "total_bits": 630519808, "q_proj": { "group_size": { @@ -39742,7 +39742,7 @@ } }, { - "accuracy": 0.9977952573252352, + "accuracy": 0.9978019853955821, "total_bits": 637526016, "q_proj": { "group_size": { @@ -39803,7 +39803,7 @@ } }, { - "accuracy": 0.9979737791183748, + "accuracy": 0.9979788443367732, "total_bits": 646987776, "q_proj": { "group_size": { @@ -39864,7 +39864,7 @@ } }, { - "accuracy": 0.9987863425753618, + "accuracy": 0.998787493984166, "total_bits": 784904192, "q_proj": { "group_size": { @@ -39925,7 +39925,7 @@ } }, { - "accuracy": 0.9989913582409683, + "accuracy": 0.9989939037906496, "total_bits": 797982720, "q_proj": { "group_size": { @@ -39986,7 +39986,7 @@ } }, { - "accuracy": 0.9991278294473886, + "accuracy": 0.9991291950603849, "total_bits": 911912960, "q_proj": { "group_size": { @@ -40038,7 +40038,7 @@ } }, { - "accuracy": 0.9994490728468487, + "accuracy": 0.9994495037550989, "total_bits": 942882816, "q_proj": { "group_size": { @@ -40090,7 +40090,7 @@ } }, { - "accuracy": 0.9997520465893965, + "accuracy": 0.9997525353131718, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -40144,7 +40144,7 @@ ], "model.layers.20.mlp": [ { - "accuracy": 0.976467280011428, + "accuracy": 0.9765426328307704, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -40196,7 +40196,7 @@ } }, { - "accuracy": 0.9770657404472953, + "accuracy": 0.9771420281184348, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -40248,7 +40248,7 @@ } }, { - "accuracy": 0.9809002342977022, + "accuracy": 0.9809642481176477, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -40297,7 +40297,7 @@ } }, { - "accuracy": 0.9821044645811382, + "accuracy": 0.9821669882849643, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -40346,7 +40346,7 @@ } }, { - "accuracy": 0.9884352338941473, + "accuracy": 0.9884727040403768, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -40398,7 +40398,7 @@ } }, { - "accuracy": 0.9893240011052081, + "accuracy": 0.9893578353681063, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -40450,7 +40450,7 @@ } }, { - "accuracy": 0.9910203681180352, + "accuracy": 0.9910467486632498, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -40499,7 +40499,7 @@ } }, { - "accuracy": 0.994028106723961, + "accuracy": 0.9940446634825907, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -40542,7 +40542,7 @@ } }, { - "accuracy": 0.9945883288195259, + "accuracy": 0.994605359670363, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -40585,7 +40585,7 @@ } }, { - "accuracy": 0.9941419978675089, + "accuracy": 0.9941594286968833, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -40637,7 +40637,7 @@ } }, { - "accuracy": 0.9948651896495568, + "accuracy": 0.9948817703284716, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -40689,7 +40689,7 @@ } }, { - "accuracy": 0.9970190713280126, + "accuracy": 0.9970267387597185, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -40741,7 +40741,7 @@ } }, { - "accuracy": 0.9974434334588679, + "accuracy": 0.9974509003131014, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -40793,7 +40793,7 @@ } }, { - "accuracy": 0.9984007692454677, + "accuracy": 0.998404477183756, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -40836,7 +40836,7 @@ } }, { - "accuracy": 0.9984816216716641, + "accuracy": 0.9984855744007387, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -40885,7 +40885,7 @@ } }, { - "accuracy": 0.9988279726944471, + "accuracy": 0.998830525988811, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -40931,7 +40931,7 @@ } }, { - "accuracy": 0.9995692382125478, + "accuracy": 0.9995703635443198, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -40973,7 +40973,7 @@ ], "model.layers.21.self_attn": [ { - "accuracy": 0.9846056668381942, + "accuracy": 0.9847216198318883, "total_bits": 320921600, "q_proj": { "group_size": { @@ -41037,7 +41037,7 @@ } }, { - "accuracy": 0.986122256831119, + "accuracy": 0.9862817667032543, "total_bits": 329244672, "q_proj": { "group_size": { @@ -41101,7 +41101,7 @@ } }, { - "accuracy": 0.9878127127885818, + "accuracy": 0.9878728013289603, "total_bits": 336188416, "q_proj": { "group_size": { @@ -41165,7 +41165,7 @@ } }, { - "accuracy": 0.9917659100733305, + "accuracy": 0.9918291600126969, "total_bits": 401721344, "q_proj": { "group_size": { @@ -41229,7 +41229,7 @@ } }, { - "accuracy": 0.9923378754603235, + "accuracy": 0.9924017285045824, "total_bits": 475443200, "q_proj": { "group_size": { @@ -41293,7 +41293,7 @@ } }, { - "accuracy": 0.9925950938149503, + "accuracy": 0.9925823674390191, "total_bits": 475642880, "q_proj": { "group_size": { @@ -41357,7 +41357,7 @@ } }, { - "accuracy": 0.9943549338924257, + "accuracy": 0.9944053841264624, "total_bits": 609923072, "q_proj": { "group_size": { @@ -41409,7 +41409,7 @@ } }, { - "accuracy": 0.9946757221692487, + "accuracy": 0.9946417267385282, "total_bits": 610188288, "q_proj": { "group_size": { @@ -41461,7 +41461,7 @@ } }, { - "accuracy": 0.9950995174677748, + "accuracy": 0.9951538290632399, "total_bits": 615184384, "q_proj": { "group_size": { @@ -41513,7 +41513,7 @@ } }, { - "accuracy": 0.9952279784177479, + "accuracy": 0.9952791705727577, "total_bits": 624115712, "q_proj": { "group_size": { @@ -41565,7 +41565,7 @@ } }, { - "accuracy": 0.9961845568920437, + "accuracy": 0.9962130079143926, "total_bits": 626637824, "q_proj": { "group_size": { @@ -41629,7 +41629,7 @@ } }, { - "accuracy": 0.9964857748464534, + "accuracy": 0.9965063000195905, "total_bits": 630519808, "q_proj": { "group_size": { @@ -41693,7 +41693,7 @@ } }, { - "accuracy": 0.9965509854649243, + "accuracy": 0.9965912248743208, "total_bits": 637526016, "q_proj": { "group_size": { @@ -41754,7 +41754,7 @@ } }, { - "accuracy": 0.9968914760178641, + "accuracy": 0.9969090337031766, "total_bits": 646987776, "q_proj": { "group_size": { @@ -41815,7 +41815,7 @@ } }, { - "accuracy": 0.9980793736482921, + "accuracy": 0.9980993808100098, "total_bits": 784904192, "q_proj": { "group_size": { @@ -41876,7 +41876,7 @@ } }, { - "accuracy": 0.9984138153101269, + "accuracy": 0.9984130559391097, "total_bits": 797982720, "q_proj": { "group_size": { @@ -41937,7 +41937,7 @@ } }, { - "accuracy": 0.9984575405128693, + "accuracy": 0.9984803692878861, "total_bits": 911912960, "q_proj": { "group_size": { @@ -41989,7 +41989,7 @@ } }, { - "accuracy": 0.9992101783128945, + "accuracy": 0.9992152751472435, "total_bits": 942882816, "q_proj": { "group_size": { @@ -42041,7 +42041,7 @@ } }, { - "accuracy": 0.9995819392467016, + "accuracy": 0.999584453964704, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -42095,7 +42095,7 @@ ], "model.layers.21.mlp": [ { - "accuracy": 0.9778591472851602, + "accuracy": 0.9779243720205206, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -42147,7 +42147,7 @@ } }, { - "accuracy": 0.9784569206990694, + "accuracy": 0.9785271271278984, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -42199,7 +42199,7 @@ } }, { - "accuracy": 0.982241911323447, + "accuracy": 0.9823044977690044, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -42248,7 +42248,7 @@ } }, { - "accuracy": 0.9834603949597007, + "accuracy": 0.9835188090801239, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -42297,7 +42297,7 @@ } }, { - "accuracy": 0.988998775419436, + "accuracy": 0.9890368643559908, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -42349,7 +42349,7 @@ } }, { - "accuracy": 0.9898957147410041, + "accuracy": 0.9899277310622366, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -42401,7 +42401,7 @@ } }, { - "accuracy": 0.9915317712645781, + "accuracy": 0.9915573557740763, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -42450,7 +42450,7 @@ } }, { - "accuracy": 0.994205222318047, + "accuracy": 0.994222417473793, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -42493,7 +42493,7 @@ } }, { - "accuracy": 0.9947732122320878, + "accuracy": 0.9947892701939532, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -42536,7 +42536,7 @@ } }, { - "accuracy": 0.9944045622097818, + "accuracy": 0.9944224581122398, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -42588,7 +42588,7 @@ } }, { - "accuracy": 0.9951068766807255, + "accuracy": 0.9951234197146014, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -42640,7 +42640,7 @@ } }, { - "accuracy": 0.9971350453009731, + "accuracy": 0.9971445612609386, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -42692,7 +42692,7 @@ } }, { - "accuracy": 0.9975504014444979, + "accuracy": 0.9975583937607313, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -42744,7 +42744,7 @@ } }, { - "accuracy": 0.9984245929670962, + "accuracy": 0.9984301101220282, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -42787,7 +42787,7 @@ } }, { - "accuracy": 0.9985393739648556, + "accuracy": 0.9985446900521454, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -42836,7 +42836,7 @@ } }, { - "accuracy": 0.9989044649624511, + "accuracy": 0.9989076275378466, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -42882,7 +42882,7 @@ } }, { - "accuracy": 0.999556878924762, + "accuracy": 0.9995569022568432, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -42924,7 +42924,7 @@ ], "model.layers.22.self_attn": [ { - "accuracy": 0.983183597263537, + "accuracy": 0.9831803974352384, "total_bits": 320921600, "q_proj": { "group_size": { @@ -42988,7 +42988,7 @@ } }, { - "accuracy": 0.9845367670059204, + "accuracy": 0.9845593963798723, "total_bits": 329244672, "q_proj": { "group_size": { @@ -43052,7 +43052,7 @@ } }, { - "accuracy": 0.9858439596075761, + "accuracy": 0.985911427359832, "total_bits": 336188416, "q_proj": { "group_size": { @@ -43116,7 +43116,7 @@ } }, { - "accuracy": 0.9902049624606183, + "accuracy": 0.9902657082206324, "total_bits": 401721344, "q_proj": { "group_size": { @@ -43180,7 +43180,7 @@ } }, { - "accuracy": 0.9917805681103155, + "accuracy": 0.9918343613022252, "total_bits": 475443200, "q_proj": { "group_size": { @@ -43244,7 +43244,7 @@ } }, { - "accuracy": 0.9919367134571075, + "accuracy": 0.9919167804090601, "total_bits": 475642880, "q_proj": { "group_size": { @@ -43308,7 +43308,7 @@ } }, { - "accuracy": 0.9944255693178428, + "accuracy": 0.9943920475871939, "total_bits": 609923072, "q_proj": { "group_size": { @@ -43360,7 +43360,7 @@ } }, { - "accuracy": 0.9946210803954225, + "accuracy": 0.994621236857615, "total_bits": 610188288, "q_proj": { "group_size": { @@ -43412,7 +43412,7 @@ } }, { - "accuracy": 0.9950337029601398, + "accuracy": 0.9950592298256723, "total_bits": 615184384, "q_proj": { "group_size": { @@ -43464,7 +43464,7 @@ } }, { - "accuracy": 0.995320494237699, + "accuracy": 0.9953598172256821, "total_bits": 624115712, "q_proj": { "group_size": { @@ -43516,7 +43516,7 @@ } }, { - "accuracy": 0.9958388174050733, + "accuracy": 0.9958708827432833, "total_bits": 626637824, "q_proj": { "group_size": { @@ -43580,7 +43580,7 @@ } }, { - "accuracy": 0.9961143763441789, + "accuracy": 0.9961241883666891, "total_bits": 630519808, "q_proj": { "group_size": { @@ -43644,7 +43644,7 @@ } }, { - "accuracy": 0.9964252851511303, + "accuracy": 0.9964320675322884, "total_bits": 637526016, "q_proj": { "group_size": { @@ -43705,7 +43705,7 @@ } }, { - "accuracy": 0.9967126642402849, + "accuracy": 0.9967340205452944, "total_bits": 646987776, "q_proj": { "group_size": { @@ -43766,7 +43766,7 @@ } }, { - "accuracy": 0.9980454113530485, + "accuracy": 0.9980517729724708, "total_bits": 784904192, "q_proj": { "group_size": { @@ -43827,7 +43827,7 @@ } }, { - "accuracy": 0.9983429463678285, + "accuracy": 0.9983548263185903, "total_bits": 797982720, "q_proj": { "group_size": { @@ -43888,7 +43888,7 @@ } }, { - "accuracy": 0.9985551038070729, + "accuracy": 0.9985572948659721, "total_bits": 911912960, "q_proj": { "group_size": { @@ -43940,7 +43940,7 @@ } }, { - "accuracy": 0.9990325910284331, + "accuracy": 0.9990374912556849, "total_bits": 942882816, "q_proj": { "group_size": { @@ -43992,7 +43992,7 @@ } }, { - "accuracy": 0.9995955505016211, + "accuracy": 0.9995979616222413, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -44046,7 +44046,7 @@ ], "model.layers.22.mlp": [ { - "accuracy": 0.9724216210214716, + "accuracy": 0.972501234004372, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -44098,7 +44098,7 @@ } }, { - "accuracy": 0.9730947331378335, + "accuracy": 0.9731665849685669, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -44150,7 +44150,7 @@ } }, { - "accuracy": 0.9774318698205446, + "accuracy": 0.9774843172023171, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -44199,7 +44199,7 @@ } }, { - "accuracy": 0.9787995956445995, + "accuracy": 0.9788479412856855, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -44248,7 +44248,7 @@ } }, { - "accuracy": 0.9861837672559839, + "accuracy": 0.9862135711469149, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -44300,7 +44300,7 @@ } }, { - "accuracy": 0.9875540701966536, + "accuracy": 0.9875905945112831, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -44352,7 +44352,7 @@ } }, { - "accuracy": 0.9894858684978987, + "accuracy": 0.9895152465293282, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -44401,7 +44401,7 @@ } }, { - "accuracy": 0.9928239811407892, + "accuracy": 0.9928451901987979, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -44444,7 +44444,7 @@ } }, { - "accuracy": 0.9936470185455523, + "accuracy": 0.9936673068686536, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -44487,7 +44487,7 @@ } }, { - "accuracy": 0.993139103839272, + "accuracy": 0.9931606148418627, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -44539,7 +44539,7 @@ } }, { - "accuracy": 0.9939876728151974, + "accuracy": 0.994006294168924, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -44591,7 +44591,7 @@ } }, { - "accuracy": 0.9964960672353443, + "accuracy": 0.996505902393868, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -44643,7 +44643,7 @@ } }, { - "accuracy": 0.9969959260993883, + "accuracy": 0.9970052320706216, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -44695,7 +44695,7 @@ } }, { - "accuracy": 0.9981006670155024, + "accuracy": 0.9981064504306567, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -44738,7 +44738,7 @@ } }, { - "accuracy": 0.9981996722911534, + "accuracy": 0.9982060919466772, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -44787,7 +44787,7 @@ } }, { - "accuracy": 0.9985992859460806, + "accuracy": 0.9986029968253876, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -44833,7 +44833,7 @@ } }, { - "accuracy": 0.9994348273857644, + "accuracy": 0.9994363433338309, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -44875,7 +44875,7 @@ ], "model.layers.23.self_attn": [ { - "accuracy": 0.9817169826281699, + "accuracy": 0.9818629732257441, "total_bits": 320921600, "q_proj": { "group_size": { @@ -44939,7 +44939,7 @@ } }, { - "accuracy": 0.9825431296699926, + "accuracy": 0.9826955026701877, "total_bits": 329244672, "q_proj": { "group_size": { @@ -45003,7 +45003,7 @@ } }, { - "accuracy": 0.9838746393981733, + "accuracy": 0.9840035124828941, "total_bits": 336188416, "q_proj": { "group_size": { @@ -45067,7 +45067,7 @@ } }, { - "accuracy": 0.9876264616062767, + "accuracy": 0.9877627562535437, "total_bits": 401721344, "q_proj": { "group_size": { @@ -45131,7 +45131,7 @@ } }, { - "accuracy": 0.9901780046914753, + "accuracy": 0.9901532115120637, "total_bits": 475443200, "q_proj": { "group_size": { @@ -45195,7 +45195,7 @@ } }, { - "accuracy": 0.9902890457918769, + "accuracy": 0.9902986939016142, "total_bits": 475642880, "q_proj": { "group_size": { @@ -45259,7 +45259,7 @@ } }, { - "accuracy": 0.9933252342437443, + "accuracy": 0.9933593524129767, "total_bits": 609923072, "q_proj": { "group_size": { @@ -45311,7 +45311,7 @@ } }, { - "accuracy": 0.9934562295675278, + "accuracy": 0.9934942467432273, "total_bits": 610188288, "q_proj": { "group_size": { @@ -45363,7 +45363,7 @@ } }, { - "accuracy": 0.9939296637710772, + "accuracy": 0.994061168479292, "total_bits": 615184384, "q_proj": { "group_size": { @@ -45415,7 +45415,7 @@ } }, { - "accuracy": 0.9941997853548903, + "accuracy": 0.9943142027447098, "total_bits": 624115712, "q_proj": { "group_size": { @@ -45467,7 +45467,7 @@ } }, { - "accuracy": 0.9941007185139155, + "accuracy": 0.994294537525428, "total_bits": 626637824, "q_proj": { "group_size": { @@ -45531,7 +45531,7 @@ } }, { - "accuracy": 0.9947139589410079, + "accuracy": 0.9946618382083742, "total_bits": 630519808, "q_proj": { "group_size": { @@ -45595,7 +45595,7 @@ } }, { - "accuracy": 0.9955674527507079, + "accuracy": 0.9955874297179674, "total_bits": 637526016, "q_proj": { "group_size": { @@ -45656,7 +45656,7 @@ } }, { - "accuracy": 0.9958606142746774, + "accuracy": 0.9958376527616852, "total_bits": 646987776, "q_proj": { "group_size": { @@ -45717,7 +45717,7 @@ } }, { - "accuracy": 0.9975546523928642, + "accuracy": 0.9976002138696218, "total_bits": 784904192, "q_proj": { "group_size": { @@ -45778,7 +45778,7 @@ } }, { - "accuracy": 0.9979352055019454, + "accuracy": 0.9979568002255339, "total_bits": 797982720, "q_proj": { "group_size": { @@ -45839,7 +45839,7 @@ } }, { - "accuracy": 0.998073333580243, + "accuracy": 0.998111776615444, "total_bits": 911912960, "q_proj": { "group_size": { @@ -45891,7 +45891,7 @@ } }, { - "accuracy": 0.9990531400238213, + "accuracy": 0.999044637146749, "total_bits": 942882816, "q_proj": { "group_size": { @@ -45943,7 +45943,7 @@ } }, { - "accuracy": 0.9994883067219665, + "accuracy": 0.999489691598635, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -45997,7 +45997,7 @@ ], "model.layers.23.mlp": [ { - "accuracy": 0.9698742879064459, + "accuracy": 0.9699594441213106, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -46049,7 +46049,7 @@ } }, { - "accuracy": 0.9706719549078691, + "accuracy": 0.9707622904526559, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -46101,7 +46101,7 @@ } }, { - "accuracy": 0.975610005228143, + "accuracy": 0.9756865893539629, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -46150,7 +46150,7 @@ } }, { - "accuracy": 0.9771576006161539, + "accuracy": 0.9772302382870725, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -46199,7 +46199,7 @@ } }, { - "accuracy": 0.9852246893079657, + "accuracy": 0.9852648236249623, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -46251,7 +46251,7 @@ } }, { - "accuracy": 0.986363244684119, + "accuracy": 0.9864037883909125, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -46303,7 +46303,7 @@ } }, { - "accuracy": 0.9885065343819166, + "accuracy": 0.9885355634124655, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -46352,7 +46352,7 @@ } }, { - "accuracy": 0.9923237413167953, + "accuracy": 0.9923412258687773, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -46395,7 +46395,7 @@ } }, { - "accuracy": 0.9930333714736136, + "accuracy": 0.9930544377941835, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -46438,7 +46438,7 @@ } }, { - "accuracy": 0.9924986440884439, + "accuracy": 0.9925215346248526, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -46490,7 +46490,7 @@ } }, { - "accuracy": 0.993418057498179, + "accuracy": 0.9934374164593848, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -46542,7 +46542,7 @@ } }, { - "accuracy": 0.9961716897393528, + "accuracy": 0.9961819484045631, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -46594,7 +46594,7 @@ } }, { - "accuracy": 0.9967142594488043, + "accuracy": 0.9967237738402266, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -46646,7 +46646,7 @@ } }, { - "accuracy": 0.9979278376620067, + "accuracy": 0.9979322776198387, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -46689,7 +46689,7 @@ } }, { - "accuracy": 0.9980486521595403, + "accuracy": 0.9980534285699066, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -46738,7 +46738,7 @@ } }, { - "accuracy": 0.9985017619634929, + "accuracy": 0.9985054923515571, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -46784,7 +46784,7 @@ } }, { - "accuracy": 0.9994325345186027, + "accuracy": 0.9994340504176522, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -46826,7 +46826,7 @@ ], "model.layers.24.self_attn": [ { - "accuracy": 0.9843216789396185, + "accuracy": 0.9843292879430872, "total_bits": 320921600, "q_proj": { "group_size": { @@ -46890,7 +46890,7 @@ } }, { - "accuracy": 0.9853914097735756, + "accuracy": 0.9854011488588232, "total_bits": 329244672, "q_proj": { "group_size": { @@ -46954,7 +46954,7 @@ } }, { - "accuracy": 0.9869614478788877, + "accuracy": 0.9869775858364607, "total_bits": 336188416, "q_proj": { "group_size": { @@ -47018,7 +47018,7 @@ } }, { - "accuracy": 0.9912432989007548, + "accuracy": 0.9912584024040323, "total_bits": 401721344, "q_proj": { "group_size": { @@ -47082,7 +47082,7 @@ } }, { - "accuracy": 0.9923934199308094, + "accuracy": 0.9924051000883705, "total_bits": 475443200, "q_proj": { "group_size": { @@ -47146,7 +47146,7 @@ } }, { - "accuracy": 0.9926053144429859, + "accuracy": 0.9926238922696364, "total_bits": 475642880, "q_proj": { "group_size": { @@ -47210,7 +47210,7 @@ } }, { - "accuracy": 0.9951531722357398, + "accuracy": 0.9951499331938592, "total_bits": 609923072, "q_proj": { "group_size": { @@ -47262,7 +47262,7 @@ } }, { - "accuracy": 0.9954400976237497, + "accuracy": 0.9954465953143019, "total_bits": 610188288, "q_proj": { "group_size": { @@ -47314,7 +47314,7 @@ } }, { - "accuracy": 0.99585509574727, + "accuracy": 0.9958527594020492, "total_bits": 615184384, "q_proj": { "group_size": { @@ -47366,7 +47366,7 @@ } }, { - "accuracy": 0.9960811145995793, + "accuracy": 0.9960865088199314, "total_bits": 624115712, "q_proj": { "group_size": { @@ -47418,7 +47418,7 @@ } }, { - "accuracy": 0.9962315528016341, + "accuracy": 0.9962407491709057, "total_bits": 626637824, "q_proj": { "group_size": { @@ -47482,7 +47482,7 @@ } }, { - "accuracy": 0.9964905800788026, + "accuracy": 0.9964871167352325, "total_bits": 630519808, "q_proj": { "group_size": { @@ -47546,7 +47546,7 @@ } }, { - "accuracy": 0.9967164503116357, + "accuracy": 0.9967130640227544, "total_bits": 637526016, "q_proj": { "group_size": { @@ -47607,7 +47607,7 @@ } }, { - "accuracy": 0.997006002029306, + "accuracy": 0.9970104174786493, "total_bits": 646987776, "q_proj": { "group_size": { @@ -47668,7 +47668,7 @@ } }, { - "accuracy": 0.9981923064118937, + "accuracy": 0.9981985570568788, "total_bits": 784904192, "q_proj": { "group_size": { @@ -47729,7 +47729,7 @@ } }, { - "accuracy": 0.9984969392810997, + "accuracy": 0.9984980035377177, "total_bits": 797982720, "q_proj": { "group_size": { @@ -47790,7 +47790,7 @@ } }, { - "accuracy": 0.9987385728837628, + "accuracy": 0.9987433651756299, "total_bits": 911912960, "q_proj": { "group_size": { @@ -47842,7 +47842,7 @@ } }, { - "accuracy": 0.9991553200310782, + "accuracy": 0.9991566678018946, "total_bits": 942882816, "q_proj": { "group_size": { @@ -47894,7 +47894,7 @@ } }, { - "accuracy": 0.9996393309219888, + "accuracy": 0.9996397776137057, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -47948,7 +47948,7 @@ ], "model.layers.24.mlp": [ { - "accuracy": 0.9686142708125868, + "accuracy": 0.9687081575393677, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -48000,7 +48000,7 @@ } }, { - "accuracy": 0.9694611210572092, + "accuracy": 0.9695503397991783, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -48052,7 +48052,7 @@ } }, { - "accuracy": 0.9742464470235925, + "accuracy": 0.974314482588517, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -48101,7 +48101,7 @@ } }, { - "accuracy": 0.9757216651188699, + "accuracy": 0.9757853918953946, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -48150,7 +48150,7 @@ } }, { - "accuracy": 0.9845417314454129, + "accuracy": 0.9845872725311079, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -48202,7 +48202,7 @@ } }, { - "accuracy": 0.9857348485996849, + "accuracy": 0.9857876661576723, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -48254,7 +48254,7 @@ } }, { - "accuracy": 0.987781654063024, + "accuracy": 0.987824841549522, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -48303,7 +48303,7 @@ } }, { - "accuracy": 0.9920061583581724, + "accuracy": 0.9920300676634437, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -48346,7 +48346,7 @@ } }, { - "accuracy": 0.9927307202627784, + "accuracy": 0.9927542374322289, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -48389,7 +48389,7 @@ } }, { - "accuracy": 0.9921597822716361, + "accuracy": 0.9921833010096299, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -48441,7 +48441,7 @@ } }, { - "accuracy": 0.9931216679121319, + "accuracy": 0.9931424377780211, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -48493,7 +48493,7 @@ } }, { - "accuracy": 0.9960059829448399, + "accuracy": 0.9960184669808337, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -48545,7 +48545,7 @@ } }, { - "accuracy": 0.9965714902469986, + "accuracy": 0.9965827665046642, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -48597,7 +48597,7 @@ } }, { - "accuracy": 0.9978459010782995, + "accuracy": 0.9978525903272001, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -48640,7 +48640,7 @@ } }, { - "accuracy": 0.9979616007522533, + "accuracy": 0.997968959769136, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -48689,7 +48689,7 @@ } }, { - "accuracy": 0.9983827977588302, + "accuracy": 0.9983889139992627, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -48735,7 +48735,7 @@ } }, { - "accuracy": 0.9994109143061858, + "accuracy": 0.9994128522904295, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -48777,7 +48777,7 @@ ], "model.layers.25.self_attn": [ { - "accuracy": 0.9843127492227053, + "accuracy": 0.9843387431220004, "total_bits": 320921600, "q_proj": { "group_size": { @@ -48841,7 +48841,7 @@ } }, { - "accuracy": 0.9851738512516022, + "accuracy": 0.985193321579381, "total_bits": 329244672, "q_proj": { "group_size": { @@ -48905,7 +48905,7 @@ } }, { - "accuracy": 0.9865232047281767, + "accuracy": 0.9865407033970481, "total_bits": 336188416, "q_proj": { "group_size": { @@ -48969,7 +48969,7 @@ } }, { - "accuracy": 0.9903055923549753, + "accuracy": 0.9903327477605719, "total_bits": 401721344, "q_proj": { "group_size": { @@ -49033,7 +49033,7 @@ } }, { - "accuracy": 0.9923056520913777, + "accuracy": 0.9923118815610283, "total_bits": 475443200, "q_proj": { "group_size": { @@ -49097,7 +49097,7 @@ } }, { - "accuracy": 0.9925013278660021, + "accuracy": 0.9925086270821722, "total_bits": 475642880, "q_proj": { "group_size": { @@ -49161,7 +49161,7 @@ } }, { - "accuracy": 0.9949398381929648, + "accuracy": 0.9949597433993691, "total_bits": 609923072, "q_proj": { "group_size": { @@ -49213,7 +49213,7 @@ } }, { - "accuracy": 0.9952043863503557, + "accuracy": 0.9952176883816719, "total_bits": 610188288, "q_proj": { "group_size": { @@ -49265,7 +49265,7 @@ } }, { - "accuracy": 0.9955900880066972, + "accuracy": 0.9956072961029253, "total_bits": 615184384, "q_proj": { "group_size": { @@ -49317,7 +49317,7 @@ } }, { - "accuracy": 0.995825889863466, + "accuracy": 0.995843663027412, "total_bits": 624115712, "q_proj": { "group_size": { @@ -49369,7 +49369,7 @@ } }, { - "accuracy": 0.9962010603202017, + "accuracy": 0.9962005764245987, "total_bits": 626637824, "q_proj": { "group_size": { @@ -49433,7 +49433,7 @@ } }, { - "accuracy": 0.9964343093727764, + "accuracy": 0.9964366382674167, "total_bits": 630519808, "q_proj": { "group_size": { @@ -49497,7 +49497,7 @@ } }, { - "accuracy": 0.9966224318272189, + "accuracy": 0.9966299106415949, "total_bits": 637526016, "q_proj": { "group_size": { @@ -49558,7 +49558,7 @@ } }, { - "accuracy": 0.996903741046002, + "accuracy": 0.9969058809311766, "total_bits": 646987776, "q_proj": { "group_size": { @@ -49619,7 +49619,7 @@ } }, { - "accuracy": 0.9981332617370706, + "accuracy": 0.9981347659700796, "total_bits": 784904192, "q_proj": { "group_size": { @@ -49680,7 +49680,7 @@ } }, { - "accuracy": 0.9984497820076189, + "accuracy": 0.9984531793743372, "total_bits": 797982720, "q_proj": { "group_size": { @@ -49741,7 +49741,7 @@ } }, { - "accuracy": 0.9986749269852513, + "accuracy": 0.998676519154718, "total_bits": 911912960, "q_proj": { "group_size": { @@ -49793,7 +49793,7 @@ } }, { - "accuracy": 0.9991186807226193, + "accuracy": 0.9991204809201392, "total_bits": 942882816, "q_proj": { "group_size": { @@ -49845,7 +49845,7 @@ } }, { - "accuracy": 0.9996262628241981, + "accuracy": 0.9996269262444816, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -49899,7 +49899,7 @@ ], "model.layers.25.mlp": [ { - "accuracy": 0.9662732387843885, + "accuracy": 0.9664000147267392, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -49951,7 +49951,7 @@ } }, { - "accuracy": 0.9671772498833506, + "accuracy": 0.9673000699595401, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -50003,7 +50003,7 @@ } }, { - "accuracy": 0.971959136034313, + "accuracy": 0.9720675129639474, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -50052,7 +50052,7 @@ } }, { - "accuracy": 0.9734430281739486, + "accuracy": 0.973544368618413, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -50101,7 +50101,7 @@ } }, { - "accuracy": 0.9832693573675657, + "accuracy": 0.9833301431254337, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -50153,7 +50153,7 @@ } }, { - "accuracy": 0.9845874278168929, + "accuracy": 0.9846476018428802, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -50205,7 +50205,7 @@ } }, { - "accuracy": 0.9866320271241037, + "accuracy": 0.9866833169209329, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -50254,7 +50254,7 @@ } }, { - "accuracy": 0.9913688349096399, + "accuracy": 0.9914020841058931, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -50297,7 +50297,7 @@ } }, { - "accuracy": 0.9921439431215587, + "accuracy": 0.9921742967869106, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -50340,7 +50340,7 @@ } }, { - "accuracy": 0.99151532508825, + "accuracy": 0.9915439455132735, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -50392,7 +50392,7 @@ } }, { - "accuracy": 0.99256458329527, + "accuracy": 0.9925924641521353, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -50444,7 +50444,7 @@ } }, { - "accuracy": 0.9956798451511484, + "accuracy": 0.9956946796492526, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -50496,7 +50496,7 @@ } }, { - "accuracy": 0.9962941063077826, + "accuracy": 0.9963079887001138, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -50548,7 +50548,7 @@ } }, { - "accuracy": 0.9976674030093771, + "accuracy": 0.9976764597782963, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -50591,7 +50591,7 @@ } }, { - "accuracy": 0.9977916435975778, + "accuracy": 0.9977993424001493, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -50640,7 +50640,7 @@ } }, { - "accuracy": 0.9982046153592435, + "accuracy": 0.99821063974186, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -50686,7 +50686,7 @@ } }, { - "accuracy": 0.9993508784100413, + "accuracy": 0.9993530133935181, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -50728,7 +50728,7 @@ ], "model.layers.26.self_attn": [ { - "accuracy": 0.9825016871878975, + "accuracy": 0.982531316970524, "total_bits": 320921600, "q_proj": { "group_size": { @@ -50792,7 +50792,7 @@ } }, { - "accuracy": 0.9834655382131275, + "accuracy": 0.983486493951396, "total_bits": 329244672, "q_proj": { "group_size": { @@ -50856,7 +50856,7 @@ } }, { - "accuracy": 0.9851887994690945, + "accuracy": 0.985211559032139, "total_bits": 336188416, "q_proj": { "group_size": { @@ -50920,7 +50920,7 @@ } }, { - "accuracy": 0.9898876524285266, + "accuracy": 0.9899178570822665, "total_bits": 401721344, "q_proj": { "group_size": { @@ -50984,7 +50984,7 @@ } }, { - "accuracy": 0.9914641364624626, + "accuracy": 0.9914805277397758, "total_bits": 475443200, "q_proj": { "group_size": { @@ -51048,7 +51048,7 @@ } }, { - "accuracy": 0.9916267575402009, + "accuracy": 0.9916451447888425, "total_bits": 475642880, "q_proj": { "group_size": { @@ -51112,7 +51112,7 @@ } }, { - "accuracy": 0.9949205717758128, + "accuracy": 0.9949221309078368, "total_bits": 609923072, "q_proj": { "group_size": { @@ -51164,7 +51164,7 @@ } }, { - "accuracy": 0.9951406482018923, + "accuracy": 0.9951507229554025, "total_bits": 610188288, "q_proj": { "group_size": { @@ -51216,7 +51216,7 @@ } }, { - "accuracy": 0.9955194608161324, + "accuracy": 0.9955254424559442, "total_bits": 615184384, "q_proj": { "group_size": { @@ -51268,7 +51268,7 @@ } }, { - "accuracy": 0.9957532392520654, + "accuracy": 0.995757990761807, "total_bits": 624115712, "q_proj": { "group_size": { @@ -51320,7 +51320,7 @@ } }, { - "accuracy": 0.995760425533119, + "accuracy": 0.9957682325651771, "total_bits": 626637824, "q_proj": { "group_size": { @@ -51384,7 +51384,7 @@ } }, { - "accuracy": 0.9960250434906859, + "accuracy": 0.9960292048360172, "total_bits": 630519808, "q_proj": { "group_size": { @@ -51448,7 +51448,7 @@ } }, { - "accuracy": 0.9963458305911014, + "accuracy": 0.9963533897933207, "total_bits": 637526016, "q_proj": { "group_size": { @@ -51509,7 +51509,7 @@ } }, { - "accuracy": 0.9966404661536217, + "accuracy": 0.9966438424430395, "total_bits": 646987776, "q_proj": { "group_size": { @@ -51570,7 +51570,7 @@ } }, { - "accuracy": 0.9979969425813148, + "accuracy": 0.9980023024897826, "total_bits": 784904192, "q_proj": { "group_size": { @@ -51631,7 +51631,7 @@ } }, { - "accuracy": 0.9983142651617527, + "accuracy": 0.998316079378128, "total_bits": 797982720, "q_proj": { "group_size": { @@ -51692,7 +51692,7 @@ } }, { - "accuracy": 0.998678772073043, + "accuracy": 0.9986804639430422, "total_bits": 911912960, "q_proj": { "group_size": { @@ -51744,7 +51744,7 @@ } }, { - "accuracy": 0.9990937805881626, + "accuracy": 0.9990934896233835, "total_bits": 942882816, "q_proj": { "group_size": { @@ -51796,7 +51796,7 @@ } }, { - "accuracy": 0.9996318916398051, + "accuracy": 0.9996323195335112, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -51850,7 +51850,7 @@ ], "model.layers.26.mlp": [ { - "accuracy": 0.9655418897929945, + "accuracy": 0.9656797490621868, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -51902,7 +51902,7 @@ } }, { - "accuracy": 0.96648896681635, + "accuracy": 0.9666231556942588, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -51954,7 +51954,7 @@ } }, { - "accuracy": 0.9711297562247828, + "accuracy": 0.9712507567907634, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -52003,7 +52003,7 @@ } }, { - "accuracy": 0.9725505207714281, + "accuracy": 0.9726638354753193, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -52052,7 +52052,7 @@ } }, { - "accuracy": 0.9828694102011228, + "accuracy": 0.9829458437467876, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -52104,7 +52104,7 @@ } }, { - "accuracy": 0.9842160720574228, + "accuracy": 0.9842789533891176, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -52156,7 +52156,7 @@ } }, { - "accuracy": 0.9861901151506525, + "accuracy": 0.9862440353945682, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -52205,7 +52205,7 @@ } }, { - "accuracy": 0.9911933211903823, + "accuracy": 0.9912319410788385, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -52248,7 +52248,7 @@ } }, { - "accuracy": 0.9919686991917459, + "accuracy": 0.9920015084116083, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -52291,7 +52291,7 @@ } }, { - "accuracy": 0.9913192281597539, + "accuracy": 0.9913559171714281, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -52343,7 +52343,7 @@ } }, { - "accuracy": 0.9923894907298841, + "accuracy": 0.9924209768834867, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -52395,7 +52395,7 @@ } }, { - "accuracy": 0.9955845824197719, + "accuracy": 0.9956043433201941, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -52447,7 +52447,7 @@ } }, { - "accuracy": 0.9962111468377867, + "accuracy": 0.9962268938359461, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -52499,7 +52499,7 @@ } }, { - "accuracy": 0.9976267936198335, + "accuracy": 0.9976374257944132, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -52542,7 +52542,7 @@ } }, { - "accuracy": 0.9977449910028985, + "accuracy": 0.9977561178568163, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -52591,7 +52591,7 @@ } }, { - "accuracy": 0.9981272024543661, + "accuracy": 0.9981363880399027, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -52637,7 +52637,7 @@ } }, { - "accuracy": 0.9993591729649588, + "accuracy": 0.999362316423733, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -52679,7 +52679,7 @@ ], "model.layers.27.self_attn": [ { - "accuracy": 0.9836736148909518, + "accuracy": 0.9836789981315011, "total_bits": 320921600, "q_proj": { "group_size": { @@ -52743,7 +52743,7 @@ } }, { - "accuracy": 0.9845029272531208, + "accuracy": 0.9845082524575686, "total_bits": 329244672, "q_proj": { "group_size": { @@ -52807,7 +52807,7 @@ } }, { - "accuracy": 0.9861994197494105, + "accuracy": 0.9862090192343059, "total_bits": 336188416, "q_proj": { "group_size": { @@ -52871,7 +52871,7 @@ } }, { - "accuracy": 0.9906061895583805, + "accuracy": 0.9906228310183475, "total_bits": 401721344, "q_proj": { "group_size": { @@ -52935,7 +52935,7 @@ } }, { - "accuracy": 0.9919237635637584, + "accuracy": 0.9919428707737672, "total_bits": 475443200, "q_proj": { "group_size": { @@ -52999,7 +52999,7 @@ } }, { - "accuracy": 0.9920995172701383, + "accuracy": 0.9921132012417442, "total_bits": 475642880, "q_proj": { "group_size": { @@ -53063,7 +53063,7 @@ } }, { - "accuracy": 0.99509250019726, + "accuracy": 0.995120180280585, "total_bits": 609923072, "q_proj": { "group_size": { @@ -53115,7 +53115,7 @@ } }, { - "accuracy": 0.9953305929114944, + "accuracy": 0.9953377999757466, "total_bits": 610188288, "q_proj": { "group_size": { @@ -53167,7 +53167,7 @@ } }, { - "accuracy": 0.9957857375082216, + "accuracy": 0.9957873731067306, "total_bits": 615184384, "q_proj": { "group_size": { @@ -53219,7 +53219,7 @@ } }, { - "accuracy": 0.9959923981835967, + "accuracy": 0.9959968893151534, "total_bits": 624115712, "q_proj": { "group_size": { @@ -53271,7 +53271,7 @@ } }, { - "accuracy": 0.9960178630916696, + "accuracy": 0.9960189516607084, "total_bits": 626637824, "q_proj": { "group_size": { @@ -53335,7 +53335,7 @@ } }, { - "accuracy": 0.9962687076706636, + "accuracy": 0.9962736372100679, "total_bits": 630519808, "q_proj": { "group_size": { @@ -53399,7 +53399,7 @@ } }, { - "accuracy": 0.9965660544602495, + "accuracy": 0.9965714871099121, "total_bits": 637526016, "q_proj": { "group_size": { @@ -53460,7 +53460,7 @@ } }, { - "accuracy": 0.9968500013806318, + "accuracy": 0.9968490882923728, "total_bits": 646987776, "q_proj": { "group_size": { @@ -53521,7 +53521,7 @@ } }, { - "accuracy": 0.9981059151652613, + "accuracy": 0.998105841247659, "total_bits": 784904192, "q_proj": { "group_size": { @@ -53582,7 +53582,7 @@ } }, { - "accuracy": 0.998419794303022, + "accuracy": 0.9984200252710205, "total_bits": 797982720, "q_proj": { "group_size": { @@ -53643,7 +53643,7 @@ } }, { - "accuracy": 0.9987193250930623, + "accuracy": 0.9987182893643254, "total_bits": 911912960, "q_proj": { "group_size": { @@ -53695,7 +53695,7 @@ } }, { - "accuracy": 0.9991493993683865, + "accuracy": 0.9991499393394119, "total_bits": 942882816, "q_proj": { "group_size": { @@ -53747,7 +53747,7 @@ } }, { - "accuracy": 0.9996352643509837, + "accuracy": 0.9996353596399882, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -53801,7 +53801,7 @@ ], "model.layers.27.mlp": [ { - "accuracy": 0.964323479878275, + "accuracy": 0.9644731032220941, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -53853,7 +53853,7 @@ } }, { - "accuracy": 0.965311398631648, + "accuracy": 0.9654506413560164, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -53905,7 +53905,7 @@ } }, { - "accuracy": 0.9699733163181105, + "accuracy": 0.9700963371678403, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -53954,7 +53954,7 @@ } }, { - "accuracy": 0.9713820281781649, + "accuracy": 0.9714985082024022, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -54003,7 +54003,7 @@ } }, { - "accuracy": 0.982238808744832, + "accuracy": 0.9823110574170163, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -54055,7 +54055,7 @@ } }, { - "accuracy": 0.9836470516104447, + "accuracy": 0.983710450561423, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -54107,7 +54107,7 @@ } }, { - "accuracy": 0.9856106623222953, + "accuracy": 0.9856660020978827, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -54156,7 +54156,7 @@ } }, { - "accuracy": 0.9908783129955593, + "accuracy": 0.990912706444138, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -54199,7 +54199,7 @@ } }, { - "accuracy": 0.991683229804039, + "accuracy": 0.9917149245738983, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -54242,7 +54242,7 @@ } }, { - "accuracy": 0.9909941773665579, + "accuracy": 0.9910297056562022, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -54294,7 +54294,7 @@ } }, { - "accuracy": 0.9921131604596188, + "accuracy": 0.992141991853714, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -54346,7 +54346,7 @@ } }, { - "accuracy": 0.9954189054275814, + "accuracy": 0.9954388396520364, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -54398,7 +54398,7 @@ } }, { - "accuracy": 0.9960752153082898, + "accuracy": 0.9960898443272239, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -54450,7 +54450,7 @@ } }, { - "accuracy": 0.9975436169065928, + "accuracy": 0.9975535453934419, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -54493,7 +54493,7 @@ } }, { - "accuracy": 0.9976613621570563, + "accuracy": 0.9976709096839553, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -54542,7 +54542,7 @@ } }, { - "accuracy": 0.9980388713118277, + "accuracy": 0.9980462922861701, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -54588,7 +54588,7 @@ } }, { - "accuracy": 0.9993408753683692, + "accuracy": 0.9993437451163405, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -54630,7 +54630,7 @@ ], "model.layers.28.self_attn": [ { - "accuracy": 0.9847634230789385, + "accuracy": 0.9847836384647771, "total_bits": 320921600, "q_proj": { "group_size": { @@ -54694,7 +54694,7 @@ } }, { - "accuracy": 0.9856616729184201, + "accuracy": 0.9856727436969155, "total_bits": 329244672, "q_proj": { "group_size": { @@ -54758,7 +54758,7 @@ } }, { - "accuracy": 0.9872836517660242, + "accuracy": 0.9873105648316836, "total_bits": 336188416, "q_proj": { "group_size": { @@ -54822,7 +54822,7 @@ } }, { - "accuracy": 0.9912071478994269, + "accuracy": 0.9912119915610865, "total_bits": 401721344, "q_proj": { "group_size": { @@ -54886,7 +54886,7 @@ } }, { - "accuracy": 0.9925311286198465, + "accuracy": 0.9925396489469629, "total_bits": 475443200, "q_proj": { "group_size": { @@ -54950,7 +54950,7 @@ } }, { - "accuracy": 0.9927132686502055, + "accuracy": 0.9927208337344622, "total_bits": 475642880, "q_proj": { "group_size": { @@ -55014,7 +55014,7 @@ } }, { - "accuracy": 0.9953283887160452, + "accuracy": 0.9953292447485422, "total_bits": 609923072, "q_proj": { "group_size": { @@ -55066,7 +55066,7 @@ } }, { - "accuracy": 0.9955696918462452, + "accuracy": 0.9955803171584481, "total_bits": 610188288, "q_proj": { "group_size": { @@ -55118,7 +55118,7 @@ } }, { - "accuracy": 0.9959665183958254, + "accuracy": 0.9959814579863298, "total_bits": 615184384, "q_proj": { "group_size": { @@ -55170,7 +55170,7 @@ } }, { - "accuracy": 0.9961686883317796, + "accuracy": 0.996179599510996, "total_bits": 624115712, "q_proj": { "group_size": { @@ -55222,7 +55222,7 @@ } }, { - "accuracy": 0.9963050061150601, + "accuracy": 0.9963118014367003, "total_bits": 626637824, "q_proj": { "group_size": { @@ -55286,7 +55286,7 @@ } }, { - "accuracy": 0.9965285619622782, + "accuracy": 0.9965381249785423, "total_bits": 630519808, "q_proj": { "group_size": { @@ -55350,7 +55350,7 @@ } }, { - "accuracy": 0.9967880527439871, + "accuracy": 0.9967977632033197, "total_bits": 637526016, "q_proj": { "group_size": { @@ -55411,7 +55411,7 @@ } }, { - "accuracy": 0.997055849550586, + "accuracy": 0.997061843934812, "total_bits": 646987776, "q_proj": { "group_size": { @@ -55472,7 +55472,7 @@ } }, { - "accuracy": 0.9982187591100994, + "accuracy": 0.9982259240197507, "total_bits": 784904192, "q_proj": { "group_size": { @@ -55533,7 +55533,7 @@ } }, { - "accuracy": 0.9985172657394096, + "accuracy": 0.9985229057308874, "total_bits": 797982720, "q_proj": { "group_size": { @@ -55594,7 +55594,7 @@ } }, { - "accuracy": 0.9987670421404274, + "accuracy": 0.998773755603715, "total_bits": 911912960, "q_proj": { "group_size": { @@ -55646,7 +55646,7 @@ } }, { - "accuracy": 0.9991831077066692, + "accuracy": 0.9991872798847524, "total_bits": 942882816, "q_proj": { "group_size": { @@ -55698,7 +55698,7 @@ } }, { - "accuracy": 0.9996453960867304, + "accuracy": 0.9996461965094664, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -55752,7 +55752,7 @@ ], "model.layers.28.mlp": [ { - "accuracy": 0.9624786878886976, + "accuracy": 0.9626379358141046, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -55804,7 +55804,7 @@ } }, { - "accuracy": 0.9634953831371508, + "accuracy": 0.9636471647965281, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -55856,7 +55856,7 @@ } }, { - "accuracy": 0.9684216631086249, + "accuracy": 0.9685488311867965, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -55905,7 +55905,7 @@ } }, { - "accuracy": 0.9699104114582664, + "accuracy": 0.9700324786336798, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -55954,7 +55954,7 @@ } }, { - "accuracy": 0.9813099440775419, + "accuracy": 0.981381157511159, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -56006,7 +56006,7 @@ } }, { - "accuracy": 0.982790071713297, + "accuracy": 0.9828622921517021, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -56058,7 +56058,7 @@ } }, { - "accuracy": 0.984847566014842, + "accuracy": 0.9849114135691994, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -56107,7 +56107,7 @@ } }, { - "accuracy": 0.9903969678439593, + "accuracy": 0.9904351132480722, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -56150,7 +56150,7 @@ } }, { - "accuracy": 0.9912386748351549, + "accuracy": 0.9912731122029456, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -56193,7 +56193,7 @@ } }, { - "accuracy": 0.9905245586445457, + "accuracy": 0.9905619503636109, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -56245,7 +56245,7 @@ } }, { - "accuracy": 0.9916995489283612, + "accuracy": 0.9917335714164534, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -56297,7 +56297,7 @@ } }, { - "accuracy": 0.9951841807679126, + "accuracy": 0.9952030809302079, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -56349,7 +56349,7 @@ } }, { - "accuracy": 0.995866101823355, + "accuracy": 0.995883831460225, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -56401,7 +56401,7 @@ } }, { - "accuracy": 0.9974081006489302, + "accuracy": 0.997419571013827, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -56444,7 +56444,7 @@ } }, { - "accuracy": 0.9975445895994964, + "accuracy": 0.9975543753489068, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -56493,7 +56493,7 @@ } }, { - "accuracy": 0.9979481826487341, + "accuracy": 0.9979561667301153, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -56539,7 +56539,7 @@ } }, { - "accuracy": 0.9993010502994845, + "accuracy": 0.9993038359833392, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -56581,7 +56581,7 @@ ], "model.layers.29.self_attn": [ { - "accuracy": 0.9815301142240825, + "accuracy": 0.9815516863998613, "total_bits": 320921600, "q_proj": { "group_size": { @@ -56645,7 +56645,7 @@ } }, { - "accuracy": 0.9823890406834451, + "accuracy": 0.9824135507407942, "total_bits": 329244672, "q_proj": { "group_size": { @@ -56709,7 +56709,7 @@ } }, { - "accuracy": 0.9843676576488897, + "accuracy": 0.9843882934043282, "total_bits": 336188416, "q_proj": { "group_size": { @@ -56773,7 +56773,7 @@ } }, { - "accuracy": 0.9891002711496855, + "accuracy": 0.9891067280581123, "total_bits": 401721344, "q_proj": { "group_size": { @@ -56837,7 +56837,7 @@ } }, { - "accuracy": 0.9909166207439021, + "accuracy": 0.9909344441012332, "total_bits": 475443200, "q_proj": { "group_size": { @@ -56901,7 +56901,7 @@ } }, { - "accuracy": 0.9910491853952408, + "accuracy": 0.9910677130285063, "total_bits": 475642880, "q_proj": { "group_size": { @@ -56965,7 +56965,7 @@ } }, { - "accuracy": 0.9947225671065482, + "accuracy": 0.994735092708939, "total_bits": 609923072, "q_proj": { "group_size": { @@ -57017,7 +57017,7 @@ } }, { - "accuracy": 0.9948854446411133, + "accuracy": 0.9948944802346983, "total_bits": 610188288, "q_proj": { "group_size": { @@ -57069,7 +57069,7 @@ } }, { - "accuracy": 0.9952379527844881, + "accuracy": 0.9952415910206343, "total_bits": 615184384, "q_proj": { "group_size": { @@ -57121,7 +57121,7 @@ } }, { - "accuracy": 0.9955002179271296, + "accuracy": 0.9955068865888997, "total_bits": 624115712, "q_proj": { "group_size": { @@ -57173,7 +57173,7 @@ } }, { - "accuracy": 0.9954644155345465, + "accuracy": 0.9954704207025076, "total_bits": 626637824, "q_proj": { "group_size": { @@ -57237,7 +57237,7 @@ } }, { - "accuracy": 0.9957557108841444, + "accuracy": 0.9957654750660846, "total_bits": 630519808, "q_proj": { "group_size": { @@ -57301,7 +57301,7 @@ } }, { - "accuracy": 0.996149329370574, + "accuracy": 0.9961572877670589, "total_bits": 637526016, "q_proj": { "group_size": { @@ -57362,7 +57362,7 @@ } }, { - "accuracy": 0.9964523460519942, + "accuracy": 0.9964592609750597, "total_bits": 646987776, "q_proj": { "group_size": { @@ -57423,7 +57423,7 @@ } }, { - "accuracy": 0.9978964207203764, + "accuracy": 0.9979009273412981, "total_bits": 784904192, "q_proj": { "group_size": { @@ -57484,7 +57484,7 @@ } }, { - "accuracy": 0.9982184461857143, + "accuracy": 0.9982216781691501, "total_bits": 797982720, "q_proj": { "group_size": { @@ -57545,7 +57545,7 @@ } }, { - "accuracy": 0.9986220072758826, + "accuracy": 0.9986260589212179, "total_bits": 911912960, "q_proj": { "group_size": { @@ -57597,7 +57597,7 @@ } }, { - "accuracy": 0.9990411081204289, + "accuracy": 0.9990428015589714, "total_bits": 942882816, "q_proj": { "group_size": { @@ -57649,7 +57649,7 @@ } }, { - "accuracy": 0.9996053627447078, + "accuracy": 0.9996058412484432, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -57703,7 +57703,7 @@ ], "model.layers.29.mlp": [ { - "accuracy": 0.9609929260454679, + "accuracy": 0.9611373418255856, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -57755,7 +57755,7 @@ } }, { - "accuracy": 0.9620737496175265, + "accuracy": 0.9622157216072083, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -57807,7 +57807,7 @@ } }, { - "accuracy": 0.9672450642836721, + "accuracy": 0.9673744502820467, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -57856,7 +57856,7 @@ } }, { - "accuracy": 0.9687830837149369, + "accuracy": 0.9689044983763444, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -57905,7 +57905,7 @@ } }, { - "accuracy": 0.9805760383605957, + "accuracy": 0.980645057402159, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -57957,7 +57957,7 @@ } }, { - "accuracy": 0.9821133566530127, + "accuracy": 0.9821816980838776, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -58009,7 +58009,7 @@ } }, { - "accuracy": 0.9842540762926403, + "accuracy": 0.9843167129315828, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -58058,7 +58058,7 @@ } }, { - "accuracy": 0.9900123124059877, + "accuracy": 0.9900516506872679, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -58101,7 +58101,7 @@ } }, { - "accuracy": 0.9908741249849922, + "accuracy": 0.9909105159734425, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -58144,7 +58144,7 @@ } }, { - "accuracy": 0.9901471428180996, + "accuracy": 0.9901822664235768, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -58196,7 +58196,7 @@ } }, { - "accuracy": 0.9913666397333145, + "accuracy": 0.9913987117378336, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -58248,7 +58248,7 @@ } }, { - "accuracy": 0.9949599582897989, + "accuracy": 0.9949776663591987, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -58300,7 +58300,7 @@ } }, { - "accuracy": 0.9956974269528138, + "accuracy": 0.9957147177897001, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -58352,7 +58352,7 @@ } }, { - "accuracy": 0.9972508937904709, + "accuracy": 0.997263913876132, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -58395,7 +58395,7 @@ } }, { - "accuracy": 0.9974375722047529, + "accuracy": 0.9974475683350312, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -58444,7 +58444,7 @@ } }, { - "accuracy": 0.997858523930374, + "accuracy": 0.9978667281960186, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -58490,7 +58490,7 @@ } }, { - "accuracy": 0.9992620605289152, + "accuracy": 0.9992656986180105, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -58532,7 +58532,7 @@ ], "model.layers.30.self_attn": [ { - "accuracy": 0.9820401213671032, + "accuracy": 0.9820604889016402, "total_bits": 320921600, "q_proj": { "group_size": { @@ -58596,7 +58596,7 @@ } }, { - "accuracy": 0.9829215156404596, + "accuracy": 0.9829492396429965, "total_bits": 329244672, "q_proj": { "group_size": { @@ -58660,7 +58660,7 @@ } }, { - "accuracy": 0.9849171811028531, + "accuracy": 0.9849240591651515, "total_bits": 336188416, "q_proj": { "group_size": { @@ -58724,7 +58724,7 @@ } }, { - "accuracy": 0.9892116071362245, + "accuracy": 0.9892054858960604, "total_bits": 401721344, "q_proj": { "group_size": { @@ -58788,7 +58788,7 @@ } }, { - "accuracy": 0.9911319393860666, + "accuracy": 0.9911259953912935, "total_bits": 475443200, "q_proj": { "group_size": { @@ -58852,7 +58852,7 @@ } }, { - "accuracy": 0.9912973603135661, + "accuracy": 0.9913017310594258, "total_bits": 475642880, "q_proj": { "group_size": { @@ -58916,7 +58916,7 @@ } }, { - "accuracy": 0.9946956242385664, + "accuracy": 0.9946931820166739, "total_bits": 609923072, "q_proj": { "group_size": { @@ -58968,7 +58968,7 @@ } }, { - "accuracy": 0.9949092986552339, + "accuracy": 0.9949109017064697, "total_bits": 610188288, "q_proj": { "group_size": { @@ -59020,7 +59020,7 @@ } }, { - "accuracy": 0.9952923208475113, + "accuracy": 0.995294920315868, "total_bits": 615184384, "q_proj": { "group_size": { @@ -59072,7 +59072,7 @@ } }, { - "accuracy": 0.9955552522289125, + "accuracy": 0.9955559051350543, "total_bits": 624115712, "q_proj": { "group_size": { @@ -59124,7 +59124,7 @@ } }, { - "accuracy": 0.995579772089657, + "accuracy": 0.9955824354761525, "total_bits": 626637824, "q_proj": { "group_size": { @@ -59188,7 +59188,7 @@ } }, { - "accuracy": 0.9958683573885968, + "accuracy": 0.9958683840538326, "total_bits": 630519808, "q_proj": { "group_size": { @@ -59252,7 +59252,7 @@ } }, { - "accuracy": 0.9962482413178996, + "accuracy": 0.9962508682357637, "total_bits": 637526016, "q_proj": { "group_size": { @@ -59313,7 +59313,7 @@ } }, { - "accuracy": 0.9965556742329347, + "accuracy": 0.9965587524991286, "total_bits": 646987776, "q_proj": { "group_size": { @@ -59374,7 +59374,7 @@ } }, { - "accuracy": 0.997934586907688, + "accuracy": 0.9979376177254476, "total_bits": 784904192, "q_proj": { "group_size": { @@ -59435,7 +59435,7 @@ } }, { - "accuracy": 0.9982653417085346, + "accuracy": 0.9982657728618697, "total_bits": 797982720, "q_proj": { "group_size": { @@ -59496,7 +59496,7 @@ } }, { - "accuracy": 0.9986030253532686, + "accuracy": 0.998604570074301, "total_bits": 911912960, "q_proj": { "group_size": { @@ -59548,7 +59548,7 @@ } }, { - "accuracy": 0.9990518495048347, + "accuracy": 0.9990519393039378, "total_bits": 942882816, "q_proj": { "group_size": { @@ -59600,7 +59600,7 @@ } }, { - "accuracy": 0.9996008325221115, + "accuracy": 0.9996023963107482, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -59654,7 +59654,7 @@ ], "model.layers.30.mlp": [ { - "accuracy": 0.9598214218491002, + "accuracy": 0.959971148716776, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -59706,7 +59706,7 @@ } }, { - "accuracy": 0.9610434927438435, + "accuracy": 0.9611950799038536, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -59758,7 +59758,7 @@ } }, { - "accuracy": 0.9664344473888999, + "accuracy": 0.9665632749858656, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -59807,7 +59807,7 @@ } }, { - "accuracy": 0.9679972968603435, + "accuracy": 0.9681248633485091, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -59856,7 +59856,7 @@ } }, { - "accuracy": 0.9800592174655512, + "accuracy": 0.9801304465846011, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -59908,7 +59908,7 @@ } }, { - "accuracy": 0.981657840703663, + "accuracy": 0.9817257601963846, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -59960,7 +59960,7 @@ } }, { - "accuracy": 0.9838469451979587, + "accuracy": 0.9839061279045908, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -60009,7 +60009,7 @@ } }, { - "accuracy": 0.9897321635170987, + "accuracy": 0.9897691183968594, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -60052,7 +60052,7 @@ } }, { - "accuracy": 0.9906195276661923, + "accuracy": 0.9906536834804636, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -60095,7 +60095,7 @@ } }, { - "accuracy": 0.9898658402656254, + "accuracy": 0.9899029676851473, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -60147,7 +60147,7 @@ } }, { - "accuracy": 0.9911341573062696, + "accuracy": 0.9911664708664543, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -60199,7 +60199,7 @@ } }, { - "accuracy": 0.9948362064988989, + "accuracy": 0.9948552137142733, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -60251,7 +60251,7 @@ } }, { - "accuracy": 0.995582335873654, + "accuracy": 0.9955986036281836, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -60303,7 +60303,7 @@ } }, { - "accuracy": 0.9972167866010415, + "accuracy": 0.9972268721382869, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -60346,7 +60346,7 @@ } }, { - "accuracy": 0.9973574845414412, + "accuracy": 0.9973660374158307, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -60395,7 +60395,7 @@ } }, { - "accuracy": 0.9977805198807466, + "accuracy": 0.9977886937558651, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -60441,7 +60441,7 @@ } }, { - "accuracy": 0.9992531056172753, + "accuracy": 0.9992555720535549, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -60483,7 +60483,7 @@ ], "model.layers.31.self_attn": [ { - "accuracy": 0.984235416901739, + "accuracy": 0.984272876852437, "total_bits": 320921600, "q_proj": { "group_size": { @@ -60547,7 +60547,7 @@ } }, { - "accuracy": 0.9849389854230379, + "accuracy": 0.9849866314938194, "total_bits": 329244672, "q_proj": { "group_size": { @@ -60611,7 +60611,7 @@ } }, { - "accuracy": 0.9865867809245461, + "accuracy": 0.9866277481380262, "total_bits": 336188416, "q_proj": { "group_size": { @@ -60675,7 +60675,7 @@ } }, { - "accuracy": 0.9900393721304441, + "accuracy": 0.9900719617542467, "total_bits": 401721344, "q_proj": { "group_size": { @@ -60739,7 +60739,7 @@ } }, { - "accuracy": 0.9922172278165817, + "accuracy": 0.9922387607787785, "total_bits": 475443200, "q_proj": { "group_size": { @@ -60803,7 +60803,7 @@ } }, { - "accuracy": 0.9923879104225259, + "accuracy": 0.9924177268618032, "total_bits": 475642880, "q_proj": { "group_size": { @@ -60867,7 +60867,7 @@ } }, { - "accuracy": 0.9952374908484911, + "accuracy": 0.9952451810240746, "total_bits": 609923072, "q_proj": { "group_size": { @@ -60919,7 +60919,7 @@ } }, { - "accuracy": 0.9954763097982657, + "accuracy": 0.995489171853191, "total_bits": 610188288, "q_proj": { "group_size": { @@ -60971,7 +60971,7 @@ } }, { - "accuracy": 0.9958613350203163, + "accuracy": 0.9958696275165206, "total_bits": 615184384, "q_proj": { "group_size": { @@ -61023,7 +61023,7 @@ } }, { - "accuracy": 0.9960536607786229, + "accuracy": 0.9960686964424033, "total_bits": 624115712, "q_proj": { "group_size": { @@ -61075,7 +61075,7 @@ } }, { - "accuracy": 0.9961375115733397, + "accuracy": 0.9961527487949321, "total_bits": 626637824, "q_proj": { "group_size": { @@ -61139,7 +61139,7 @@ } }, { - "accuracy": 0.9963771587139681, + "accuracy": 0.9963879400962278, "total_bits": 630519808, "q_proj": { "group_size": { @@ -61203,7 +61203,7 @@ } }, { - "accuracy": 0.9966729044129974, + "accuracy": 0.9966846194706465, "total_bits": 637526016, "q_proj": { "group_size": { @@ -61264,7 +61264,7 @@ } }, { - "accuracy": 0.996945564488047, + "accuracy": 0.996952966835938, "total_bits": 646987776, "q_proj": { "group_size": { @@ -61325,7 +61325,7 @@ } }, { - "accuracy": 0.9981675210751986, + "accuracy": 0.9981722025886962, "total_bits": 784904192, "q_proj": { "group_size": { @@ -61386,7 +61386,7 @@ } }, { - "accuracy": 0.9984648233573687, + "accuracy": 0.998467809079509, "total_bits": 797982720, "q_proj": { "group_size": { @@ -61447,7 +61447,7 @@ } }, { - "accuracy": 0.9987511289747137, + "accuracy": 0.9987523518502712, "total_bits": 911912960, "q_proj": { "group_size": { @@ -61499,7 +61499,7 @@ } }, { - "accuracy": 0.9991459296526093, + "accuracy": 0.9991487039155081, "total_bits": 942882816, "q_proj": { "group_size": { @@ -61551,7 +61551,7 @@ } }, { - "accuracy": 0.9996340804684319, + "accuracy": 0.99963509509536, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -61605,7 +61605,7 @@ ], "model.layers.31.mlp": [ { - "accuracy": 0.9587835826371846, + "accuracy": 0.9589311229555231, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -61657,7 +61657,7 @@ } }, { - "accuracy": 0.9600702304589122, + "accuracy": 0.9602131341633043, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -61709,7 +61709,7 @@ } }, { - "accuracy": 0.9654311224033958, + "accuracy": 0.9655649944355613, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -61758,7 +61758,7 @@ } }, { - "accuracy": 0.9669944393007379, + "accuracy": 0.9671243962488676, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -61807,7 +61807,7 @@ } }, { - "accuracy": 0.9795444592049247, + "accuracy": 0.979615906351491, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -61859,7 +61859,7 @@ } }, { - "accuracy": 0.9811756359903436, + "accuracy": 0.9812428590498472, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -61911,7 +61911,7 @@ } }, { - "accuracy": 0.9833624221776661, + "accuracy": 0.9834234541968295, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -61960,7 +61960,7 @@ } }, { - "accuracy": 0.9894664695388392, + "accuracy": 0.9895042596678985, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -62003,7 +62003,7 @@ } }, { - "accuracy": 0.9903580687547985, + "accuracy": 0.9903908387610787, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -62046,7 +62046,7 @@ } }, { - "accuracy": 0.9896129110926076, + "accuracy": 0.9896476896185624, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -62098,7 +62098,7 @@ } }, { - "accuracy": 0.990903830841968, + "accuracy": 0.9909345288025705, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -62150,7 +62150,7 @@ } }, { - "accuracy": 0.9947126703827005, + "accuracy": 0.9947292640020973, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -62202,7 +62202,7 @@ } }, { - "accuracy": 0.9954664036631584, + "accuracy": 0.9954804636930165, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -62254,7 +62254,7 @@ } }, { - "accuracy": 0.9971409316517805, + "accuracy": 0.9971516316658572, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -62297,7 +62297,7 @@ } }, { - "accuracy": 0.9972920727572943, + "accuracy": 0.99730234593153, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -62346,7 +62346,7 @@ } }, { - "accuracy": 0.9977114161378459, + "accuracy": 0.9977199888150943, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -62392,7 +62392,7 @@ } }, { - "accuracy": 0.9992278917251449, + "accuracy": 0.9992309948429465, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -62434,7 +62434,7 @@ ], "model.layers.32.self_attn": [ { - "accuracy": 0.983788571859661, + "accuracy": 0.9837789315926401, "total_bits": 320921600, "q_proj": { "group_size": { @@ -62498,7 +62498,7 @@ } }, { - "accuracy": 0.9846030065887853, + "accuracy": 0.9845799521396035, "total_bits": 329244672, "q_proj": { "group_size": { @@ -62562,7 +62562,7 @@ } }, { - "accuracy": 0.9862108763895536, + "accuracy": 0.9861935298693808, "total_bits": 336188416, "q_proj": { "group_size": { @@ -62626,7 +62626,7 @@ } }, { - "accuracy": 0.9899823391123822, + "accuracy": 0.9899732172489166, "total_bits": 401721344, "q_proj": { "group_size": { @@ -62690,7 +62690,7 @@ } }, { - "accuracy": 0.9920371096385153, + "accuracy": 0.9920279571884557, "total_bits": 475443200, "q_proj": { "group_size": { @@ -62754,7 +62754,7 @@ } }, { - "accuracy": 0.9921974335846148, + "accuracy": 0.9921987982172715, "total_bits": 475642880, "q_proj": { "group_size": { @@ -62818,7 +62818,7 @@ } }, { - "accuracy": 0.9950885431546914, + "accuracy": 0.9950730722201498, "total_bits": 609923072, "q_proj": { "group_size": { @@ -62870,7 +62870,7 @@ } }, { - "accuracy": 0.9953069561406186, + "accuracy": 0.9953134079512796, "total_bits": 610188288, "q_proj": { "group_size": { @@ -62922,7 +62922,7 @@ } }, { - "accuracy": 0.9957514895420325, + "accuracy": 0.9957453839873013, "total_bits": 615184384, "q_proj": { "group_size": { @@ -62974,7 +62974,7 @@ } }, { - "accuracy": 0.9959852922903863, + "accuracy": 0.995971751840491, "total_bits": 624115712, "q_proj": { "group_size": { @@ -63026,7 +63026,7 @@ } }, { - "accuracy": 0.9960485390926662, + "accuracy": 0.9960439977677245, "total_bits": 626637824, "q_proj": { "group_size": { @@ -63090,7 +63090,7 @@ } }, { - "accuracy": 0.9963084474990243, + "accuracy": 0.9962991472137602, "total_bits": 630519808, "q_proj": { "group_size": { @@ -63154,7 +63154,7 @@ } }, { - "accuracy": 0.9965534484700153, + "accuracy": 0.9965502443282228, "total_bits": 637526016, "q_proj": { "group_size": { @@ -63215,7 +63215,7 @@ } }, { - "accuracy": 0.996860356707322, + "accuracy": 0.9968534704101714, "total_bits": 646987776, "q_proj": { "group_size": { @@ -63276,7 +63276,7 @@ } }, { - "accuracy": 0.99807982048706, + "accuracy": 0.9980759375581616, "total_bits": 784904192, "q_proj": { "group_size": { @@ -63337,7 +63337,7 @@ } }, { - "accuracy": 0.9984136358099548, + "accuracy": 0.9984066064812636, "total_bits": 797982720, "q_proj": { "group_size": { @@ -63398,7 +63398,7 @@ } }, { - "accuracy": 0.9986894812042776, + "accuracy": 0.9986898150098952, "total_bits": 911912960, "q_proj": { "group_size": { @@ -63450,7 +63450,7 @@ } }, { - "accuracy": 0.9991304532281662, + "accuracy": 0.9991280624741002, "total_bits": 942882816, "q_proj": { "group_size": { @@ -63502,7 +63502,7 @@ } }, { - "accuracy": 0.9996179142770799, + "accuracy": 0.9996176984063104, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -63556,7 +63556,7 @@ ], "model.layers.32.mlp": [ { - "accuracy": 0.9586603233688756, + "accuracy": 0.9588103168889096, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -63608,7 +63608,7 @@ } }, { - "accuracy": 0.959919622069911, + "accuracy": 0.9600704061357599, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -63660,7 +63660,7 @@ } }, { - "accuracy": 0.9652071877529746, + "accuracy": 0.9653371070560656, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -63709,7 +63709,7 @@ } }, { - "accuracy": 0.9667495489120483, + "accuracy": 0.9668755123489782, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -63758,7 +63758,7 @@ } }, { - "accuracy": 0.979425930663159, + "accuracy": 0.979493299597188, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -63810,7 +63810,7 @@ } }, { - "accuracy": 0.9810855247472462, + "accuracy": 0.9811552668872633, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -63862,7 +63862,7 @@ } }, { - "accuracy": 0.9832430042718586, + "accuracy": 0.9833088360334697, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -63911,7 +63911,7 @@ } }, { - "accuracy": 0.9893737756892255, + "accuracy": 0.9894092780979056, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -63954,7 +63954,7 @@ } }, { - "accuracy": 0.9902974124017515, + "accuracy": 0.9903335539918197, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -63997,7 +63997,7 @@ } }, { - "accuracy": 0.9895429124957636, + "accuracy": 0.9895738708345514, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -64049,7 +64049,7 @@ } }, { - "accuracy": 0.990845359469715, + "accuracy": 0.9908769005223325, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -64101,7 +64101,7 @@ } }, { - "accuracy": 0.9946664485492205, + "accuracy": 0.9946769256340829, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -64153,7 +64153,7 @@ } }, { - "accuracy": 0.9954248125615873, + "accuracy": 0.9954393121757006, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -64205,7 +64205,7 @@ } }, { - "accuracy": 0.9971077991159338, + "accuracy": 0.9971106103376338, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -64248,7 +64248,7 @@ } }, { - "accuracy": 0.9972676977907357, + "accuracy": 0.9972736047286737, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -64297,7 +64297,7 @@ } }, { - "accuracy": 0.9976855630153104, + "accuracy": 0.9976899345454416, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -64343,7 +64343,7 @@ } }, { - "accuracy": 0.9991996156934061, + "accuracy": 0.9992015682356922, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -64385,7 +64385,7 @@ ], "model.layers.33.self_attn": [ { - "accuracy": 0.9832363787450289, + "accuracy": 0.9832709353221091, "total_bits": 320921600, "q_proj": { "group_size": { @@ -64449,7 +64449,7 @@ } }, { - "accuracy": 0.983958729003605, + "accuracy": 0.9840328552220997, "total_bits": 329244672, "q_proj": { "group_size": { @@ -64513,7 +64513,7 @@ } }, { - "accuracy": 0.9858537880997909, + "accuracy": 0.9858949733407874, "total_bits": 336188416, "q_proj": { "group_size": { @@ -64577,7 +64577,7 @@ } }, { - "accuracy": 0.98956876365762, + "accuracy": 0.9895585304812381, "total_bits": 401721344, "q_proj": { "group_size": { @@ -64641,7 +64641,7 @@ } }, { - "accuracy": 0.9916938974669105, + "accuracy": 0.9916910372282329, "total_bits": 475443200, "q_proj": { "group_size": { @@ -64705,7 +64705,7 @@ } }, { - "accuracy": 0.9919715288438296, + "accuracy": 0.991985526524092, "total_bits": 475642880, "q_proj": { "group_size": { @@ -64769,7 +64769,7 @@ } }, { - "accuracy": 0.9948874962957281, + "accuracy": 0.9948623941132897, "total_bits": 609923072, "q_proj": { "group_size": { @@ -64821,7 +64821,7 @@ } }, { - "accuracy": 0.9952712031571489, + "accuracy": 0.995282561371201, "total_bits": 610188288, "q_proj": { "group_size": { @@ -64873,7 +64873,7 @@ } }, { - "accuracy": 0.9957110207331809, + "accuracy": 0.9957207355060076, "total_bits": 615184384, "q_proj": { "group_size": { @@ -64925,7 +64925,7 @@ } }, { - "accuracy": 0.9959578306267136, + "accuracy": 0.9959697539084836, "total_bits": 624115712, "q_proj": { "group_size": { @@ -64977,7 +64977,7 @@ } }, { - "accuracy": 0.9959167249892887, + "accuracy": 0.9959362070811423, "total_bits": 626637824, "q_proj": { "group_size": { @@ -65041,7 +65041,7 @@ } }, { - "accuracy": 0.9962116393603777, + "accuracy": 0.9962208694533298, "total_bits": 630519808, "q_proj": { "group_size": { @@ -65105,7 +65105,7 @@ } }, { - "accuracy": 0.9965639149672106, + "accuracy": 0.996567261846442, "total_bits": 637526016, "q_proj": { "group_size": { @@ -65166,7 +65166,7 @@ } }, { - "accuracy": 0.9968710359382001, + "accuracy": 0.9968753521771807, "total_bits": 646987776, "q_proj": { "group_size": { @@ -65227,7 +65227,7 @@ } }, { - "accuracy": 0.9980409345344493, + "accuracy": 0.9980408598325754, "total_bits": 784904192, "q_proj": { "group_size": { @@ -65288,7 +65288,7 @@ } }, { - "accuracy": 0.9984059478891524, + "accuracy": 0.9984064983498109, "total_bits": 797982720, "q_proj": { "group_size": { @@ -65349,7 +65349,7 @@ } }, { - "accuracy": 0.9986459696175236, + "accuracy": 0.9986430949678546, "total_bits": 911912960, "q_proj": { "group_size": { @@ -65401,7 +65401,7 @@ } }, { - "accuracy": 0.9991089876152968, + "accuracy": 0.9991138013766, "total_bits": 942882816, "q_proj": { "group_size": { @@ -65453,7 +65453,7 @@ } }, { - "accuracy": 0.9995804813837534, + "accuracy": 0.9995791603271899, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -65507,7 +65507,7 @@ ], "model.layers.33.mlp": [ { - "accuracy": 0.9581893149175142, + "accuracy": 0.9583347220169871, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -65559,7 +65559,7 @@ } }, { - "accuracy": 0.9595503399246618, + "accuracy": 0.9596848958416989, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -65611,7 +65611,7 @@ } }, { - "accuracy": 0.965007273774398, + "accuracy": 0.9651185180011549, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -65660,7 +65660,7 @@ } }, { - "accuracy": 0.9665590148223073, + "accuracy": 0.9666644899468673, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -65709,7 +65709,7 @@ } }, { - "accuracy": 0.9792844496275249, + "accuracy": 0.9793496916168615, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -65761,7 +65761,7 @@ } }, { - "accuracy": 0.9809349147897017, + "accuracy": 0.9809939892668473, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -65813,7 +65813,7 @@ } }, { - "accuracy": 0.9831100012126722, + "accuracy": 0.9831628752382178, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -65862,7 +65862,7 @@ } }, { - "accuracy": 0.9893192837112829, + "accuracy": 0.9893531579720346, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -65905,7 +65905,7 @@ } }, { - "accuracy": 0.9902216564667853, + "accuracy": 0.9902546499904833, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -65948,7 +65948,7 @@ } }, { - "accuracy": 0.9894829847310719, + "accuracy": 0.9895142481515282, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -66000,7 +66000,7 @@ } }, { - "accuracy": 0.9907823393219396, + "accuracy": 0.9908116295149452, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -66052,7 +66052,7 @@ } }, { - "accuracy": 0.9946439779902759, + "accuracy": 0.9946610370748922, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -66104,7 +66104,7 @@ } }, { - "accuracy": 0.9954071684103263, + "accuracy": 0.9954227675732813, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -66156,7 +66156,7 @@ } }, { - "accuracy": 0.9971053355226391, + "accuracy": 0.9971146087505316, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -66199,7 +66199,7 @@ } }, { - "accuracy": 0.997255968616197, + "accuracy": 0.9972661900285044, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -66248,7 +66248,7 @@ } }, { - "accuracy": 0.9976842087742529, + "accuracy": 0.9976926877310401, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -66294,7 +66294,7 @@ } }, { - "accuracy": 0.9992150284938122, + "accuracy": 0.9992178338334748, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -66336,7 +66336,7 @@ ], "model.layers.34.self_attn": [ { - "accuracy": 0.9854075469468769, + "accuracy": 0.9855087242628399, "total_bits": 320921600, "q_proj": { "group_size": { @@ -66400,7 +66400,7 @@ } }, { - "accuracy": 0.9860877143709283, + "accuracy": 0.9861175904148504, "total_bits": 329244672, "q_proj": { "group_size": { @@ -66464,7 +66464,7 @@ } }, { - "accuracy": 0.987880247988199, + "accuracy": 0.9879126909532046, "total_bits": 336188416, "q_proj": { "group_size": { @@ -66528,7 +66528,7 @@ } }, { - "accuracy": 0.9905434831192619, + "accuracy": 0.9905880752362703, "total_bits": 401721344, "q_proj": { "group_size": { @@ -66592,7 +66592,7 @@ } }, { - "accuracy": 0.9927661175790586, + "accuracy": 0.9927895555370733, "total_bits": 475443200, "q_proj": { "group_size": { @@ -66656,7 +66656,7 @@ } }, { - "accuracy": 0.9929785344161486, + "accuracy": 0.9930031942693811, "total_bits": 475642880, "q_proj": { "group_size": { @@ -66720,7 +66720,7 @@ } }, { - "accuracy": 0.9954419967375303, + "accuracy": 0.9954403615311572, "total_bits": 609923072, "q_proj": { "group_size": { @@ -66772,7 +66772,7 @@ } }, { - "accuracy": 0.995711609329048, + "accuracy": 0.9957316478616313, "total_bits": 610188288, "q_proj": { "group_size": { @@ -66824,7 +66824,7 @@ } }, { - "accuracy": 0.9962127922396911, + "accuracy": 0.9962147670356851, "total_bits": 615184384, "q_proj": { "group_size": { @@ -66876,7 +66876,7 @@ } }, { - "accuracy": 0.9964091911127693, + "accuracy": 0.9964160083940155, "total_bits": 624115712, "q_proj": { "group_size": { @@ -66928,7 +66928,7 @@ } }, { - "accuracy": 0.996440598054936, + "accuracy": 0.9964508327998614, "total_bits": 626637824, "q_proj": { "group_size": { @@ -66992,7 +66992,7 @@ } }, { - "accuracy": 0.9967069884664134, + "accuracy": 0.9967052411091956, "total_bits": 630519808, "q_proj": { "group_size": { @@ -67056,7 +67056,7 @@ } }, { - "accuracy": 0.9969391160105404, + "accuracy": 0.9969382058633002, "total_bits": 637526016, "q_proj": { "group_size": { @@ -67117,7 +67117,7 @@ } }, { - "accuracy": 0.9972438873036912, + "accuracy": 0.9972501250081941, "total_bits": 646987776, "q_proj": { "group_size": { @@ -67178,7 +67178,7 @@ } }, { - "accuracy": 0.9982861241227702, + "accuracy": 0.9982923831986753, "total_bits": 784904192, "q_proj": { "group_size": { @@ -67239,7 +67239,7 @@ } }, { - "accuracy": 0.9985934856690859, + "accuracy": 0.9985986613717518, "total_bits": 797982720, "q_proj": { "group_size": { @@ -67300,7 +67300,7 @@ } }, { - "accuracy": 0.9987810950138067, + "accuracy": 0.9987855917332988, "total_bits": 911912960, "q_proj": { "group_size": { @@ -67352,7 +67352,7 @@ } }, { - "accuracy": 0.9992086080049998, + "accuracy": 0.9992090294039563, "total_bits": 942882816, "q_proj": { "group_size": { @@ -67404,7 +67404,7 @@ } }, { - "accuracy": 0.9996133791772943, + "accuracy": 0.9996135915188413, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -67458,7 +67458,7 @@ ], "model.layers.34.mlp": [ { - "accuracy": 0.9577239908670124, + "accuracy": 0.9578617591606944, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -67510,7 +67510,7 @@ } }, { - "accuracy": 0.9590804670986376, + "accuracy": 0.959228841882003, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -67562,7 +67562,7 @@ } }, { - "accuracy": 0.9645362308150843, + "accuracy": 0.9646659838525873, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -67611,7 +67611,7 @@ } }, { - "accuracy": 0.9660817039640326, + "accuracy": 0.966205706721858, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -67660,7 +67660,7 @@ } }, { - "accuracy": 0.979001344818818, + "accuracy": 0.9790830831778677, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -67712,7 +67712,7 @@ } }, { - "accuracy": 0.9806966091457167, + "accuracy": 0.9807654979981875, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -67764,7 +67764,7 @@ } }, { - "accuracy": 0.9828820730510511, + "accuracy": 0.9829427364625429, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -67813,7 +67813,7 @@ } }, { - "accuracy": 0.9891855685334456, + "accuracy": 0.9892268212218034, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -67856,7 +67856,7 @@ } }, { - "accuracy": 0.9901027083396912, + "accuracy": 0.9901393353939056, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -67899,7 +67899,7 @@ } }, { - "accuracy": 0.9893196413391515, + "accuracy": 0.989357211872151, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -67951,7 +67951,7 @@ } }, { - "accuracy": 0.9906583389169291, + "accuracy": 0.9906921127909109, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -68003,7 +68003,7 @@ } }, { - "accuracy": 0.9945507335819697, + "accuracy": 0.9945739135146141, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -68055,7 +68055,7 @@ } }, { - "accuracy": 0.9953426028552809, + "accuracy": 0.9953595470441016, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -68107,7 +68107,7 @@ } }, { - "accuracy": 0.9970535312436128, + "accuracy": 0.9970654425652403, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -68150,7 +68150,7 @@ } }, { - "accuracy": 0.9972035051568559, + "accuracy": 0.9972154109885818, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -68199,7 +68199,7 @@ } }, { - "accuracy": 0.9976178652753955, + "accuracy": 0.9976286409716857, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -68245,7 +68245,7 @@ } }, { - "accuracy": 0.9991879999931705, + "accuracy": 0.9991912010959104, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -68287,7 +68287,7 @@ ], "model.layers.35.self_attn": [ { - "accuracy": 0.9854714619485956, + "accuracy": 0.9855219721794128, "total_bits": 320921600, "q_proj": { "group_size": { @@ -68351,7 +68351,7 @@ } }, { - "accuracy": 0.9861034076464804, + "accuracy": 0.9861412064025277, "total_bits": 329244672, "q_proj": { "group_size": { @@ -68415,7 +68415,7 @@ } }, { - "accuracy": 0.9879133920920523, + "accuracy": 0.9879899628852543, "total_bits": 336188416, "q_proj": { "group_size": { @@ -68479,7 +68479,7 @@ } }, { - "accuracy": 0.9907108545303345, + "accuracy": 0.9907715948004472, "total_bits": 401721344, "q_proj": { "group_size": { @@ -68543,7 +68543,7 @@ } }, { - "accuracy": 0.9927917734572762, + "accuracy": 0.9928295510379892, "total_bits": 475443200, "q_proj": { "group_size": { @@ -68607,7 +68607,7 @@ } }, { - "accuracy": 0.9930278039292285, + "accuracy": 0.993051299139073, "total_bits": 475642880, "q_proj": { "group_size": { @@ -68671,7 +68671,7 @@ } }, { - "accuracy": 0.9953333398229197, + "accuracy": 0.9953560884061613, "total_bits": 609923072, "q_proj": { "group_size": { @@ -68723,7 +68723,7 @@ } }, { - "accuracy": 0.995655786834265, + "accuracy": 0.9956709644512126, "total_bits": 610188288, "q_proj": { "group_size": { @@ -68775,7 +68775,7 @@ } }, { - "accuracy": 0.9961851674475168, + "accuracy": 0.9961804159377751, "total_bits": 615184384, "q_proj": { "group_size": { @@ -68827,7 +68827,7 @@ } }, { - "accuracy": 0.9963640335359072, + "accuracy": 0.9963702653583727, "total_bits": 624115712, "q_proj": { "group_size": { @@ -68879,7 +68879,7 @@ } }, { - "accuracy": 0.9964556682266688, + "accuracy": 0.9964685796907073, "total_bits": 626637824, "q_proj": { "group_size": { @@ -68943,7 +68943,7 @@ } }, { - "accuracy": 0.9967677446180269, + "accuracy": 0.9967787759868723, "total_bits": 630519808, "q_proj": { "group_size": { @@ -69007,7 +69007,7 @@ } }, { - "accuracy": 0.9969205372035503, + "accuracy": 0.9969335112132525, "total_bits": 637526016, "q_proj": { "group_size": { @@ -69068,7 +69068,7 @@ } }, { - "accuracy": 0.9972705262664118, + "accuracy": 0.9972797702801856, "total_bits": 646987776, "q_proj": { "group_size": { @@ -69129,7 +69129,7 @@ } }, { - "accuracy": 0.998319965444113, + "accuracy": 0.9983225760883406, "total_bits": 784904192, "q_proj": { "group_size": { @@ -69190,7 +69190,7 @@ } }, { - "accuracy": 0.9986249983898903, + "accuracy": 0.9986316860702477, "total_bits": 797982720, "q_proj": { "group_size": { @@ -69251,7 +69251,7 @@ } }, { - "accuracy": 0.9988023847537605, + "accuracy": 0.9988022412320501, "total_bits": 911912960, "q_proj": { "group_size": { @@ -69303,7 +69303,7 @@ } }, { - "accuracy": 0.9992577754647324, + "accuracy": 0.9992609944586691, "total_bits": 942882816, "q_proj": { "group_size": { @@ -69355,7 +69355,7 @@ } }, { - "accuracy": 0.999637408083991, + "accuracy": 0.9996381553478146, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -69409,7 +69409,7 @@ ], "model.layers.35.mlp": [ { - "accuracy": 0.9571282644020884, + "accuracy": 0.9572559908816689, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -69461,7 +69461,7 @@ } }, { - "accuracy": 0.9584953628088299, + "accuracy": 0.958608257143121, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -69513,7 +69513,7 @@ } }, { - "accuracy": 0.9639712823064703, + "accuracy": 0.9640724627595199, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -69562,7 +69562,7 @@ } }, { - "accuracy": 0.9655094899629292, + "accuracy": 0.9656036778500205, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -69611,7 +69611,7 @@ } }, { - "accuracy": 0.9786689736341175, + "accuracy": 0.9787614659259194, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -69663,7 +69663,7 @@ } }, { - "accuracy": 0.9804144878136484, + "accuracy": 0.9804871599925192, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -69715,7 +69715,7 @@ } }, { - "accuracy": 0.9826081238294903, + "accuracy": 0.9826749249508506, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -69764,7 +69764,7 @@ } }, { - "accuracy": 0.9890577251973905, + "accuracy": 0.9890931483946348, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -69807,7 +69807,7 @@ } }, { - "accuracy": 0.989974214842445, + "accuracy": 0.9900113391248804, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -69850,7 +69850,7 @@ } }, { - "accuracy": 0.9891672424579921, + "accuracy": 0.9892040796970066, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -69902,7 +69902,7 @@ } }, { - "accuracy": 0.9905268455806532, + "accuracy": 0.9905607237627632, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -69954,7 +69954,7 @@ } }, { - "accuracy": 0.9944821407920436, + "accuracy": 0.9944987795070598, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -70006,7 +70006,7 @@ } }, { - "accuracy": 0.9952786827557966, + "accuracy": 0.9952948536527785, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -70058,7 +70058,7 @@ } }, { - "accuracy": 0.9970202077376215, + "accuracy": 0.9970290090300535, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -70101,7 +70101,7 @@ } }, { - "accuracy": 0.997175001588307, + "accuracy": 0.9971843344208441, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -70150,7 +70150,7 @@ } }, { - "accuracy": 0.9975908433920458, + "accuracy": 0.9975983027956987, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -70196,7 +70196,7 @@ } }, { - "accuracy": 0.999188064695581, + "accuracy": 0.9991906096080416, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -70238,7 +70238,7 @@ ], "model.layers.36.self_attn": [ { - "accuracy": 0.9863045011696062, + "accuracy": 0.9863390451983401, "total_bits": 320921600, "q_proj": { "group_size": { @@ -70302,7 +70302,7 @@ } }, { - "accuracy": 0.986893438979199, + "accuracy": 0.9868774484646948, "total_bits": 329244672, "q_proj": { "group_size": { @@ -70366,7 +70366,7 @@ } }, { - "accuracy": 0.9885617463212264, + "accuracy": 0.98854960501194, "total_bits": 336188416, "q_proj": { "group_size": { @@ -70430,7 +70430,7 @@ } }, { - "accuracy": 0.9913396811798999, + "accuracy": 0.9913351810292194, "total_bits": 401721344, "q_proj": { "group_size": { @@ -70494,7 +70494,7 @@ } }, { - "accuracy": 0.9931834591062445, + "accuracy": 0.9931937448288265, "total_bits": 475443200, "q_proj": { "group_size": { @@ -70558,7 +70558,7 @@ } }, { - "accuracy": 0.9933706569044214, + "accuracy": 0.9933542758226395, "total_bits": 475642880, "q_proj": { "group_size": { @@ -70622,7 +70622,7 @@ } }, { - "accuracy": 0.9958024981774782, + "accuracy": 0.9957951809230604, "total_bits": 609923072, "q_proj": { "group_size": { @@ -70674,7 +70674,7 @@ } }, { - "accuracy": 0.996031147084738, + "accuracy": 0.9960473983695632, "total_bits": 610188288, "q_proj": { "group_size": { @@ -70726,7 +70726,7 @@ } }, { - "accuracy": 0.9964288124128392, + "accuracy": 0.9964395463466644, "total_bits": 615184384, "q_proj": { "group_size": { @@ -70778,7 +70778,7 @@ } }, { - "accuracy": 0.9966050135461908, + "accuracy": 0.9966199441175712, "total_bits": 624115712, "q_proj": { "group_size": { @@ -70830,7 +70830,7 @@ } }, { - "accuracy": 0.9966244474053383, + "accuracy": 0.9966326367698217, "total_bits": 626637824, "q_proj": { "group_size": { @@ -70894,7 +70894,7 @@ } }, { - "accuracy": 0.996872601736533, + "accuracy": 0.9968824076809382, "total_bits": 630519808, "q_proj": { "group_size": { @@ -70958,7 +70958,7 @@ } }, { - "accuracy": 0.9971074558010227, + "accuracy": 0.9971085827993719, "total_bits": 637526016, "q_proj": { "group_size": { @@ -71019,7 +71019,7 @@ } }, { - "accuracy": 0.9973785700766664, + "accuracy": 0.9973862284892484, "total_bits": 646987776, "q_proj": { "group_size": { @@ -71080,7 +71080,7 @@ } }, { - "accuracy": 0.998377767048384, + "accuracy": 0.9983847175577754, "total_bits": 784904192, "q_proj": { "group_size": { @@ -71141,7 +71141,7 @@ } }, { - "accuracy": 0.9986718816584662, + "accuracy": 0.9986747703269908, "total_bits": 797982720, "q_proj": { "group_size": { @@ -71202,7 +71202,7 @@ } }, { - "accuracy": 0.9988700979830403, + "accuracy": 0.9988749993866995, "total_bits": 911912960, "q_proj": { "group_size": { @@ -71254,7 +71254,7 @@ } }, { - "accuracy": 0.9992914740976534, + "accuracy": 0.9992924747792514, "total_bits": 942882816, "q_proj": { "group_size": { @@ -71306,7 +71306,7 @@ } }, { - "accuracy": 0.9996670260780344, + "accuracy": 0.9996655378490686, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -71360,7 +71360,7 @@ ], "model.layers.36.mlp": [ { - "accuracy": 0.9572205637630663, + "accuracy": 0.9573487702168917, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -71412,7 +71412,7 @@ } }, { - "accuracy": 0.958497166633606, + "accuracy": 0.9586402962082311, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -71464,7 +71464,7 @@ } }, { - "accuracy": 0.9639960182340521, + "accuracy": 0.9641141577770835, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -71513,7 +71513,7 @@ } }, { - "accuracy": 0.9655340941328752, + "accuracy": 0.965648983654223, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -71562,7 +71562,7 @@ } }, { - "accuracy": 0.9786880737856815, + "accuracy": 0.9787535181171015, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -71614,7 +71614,7 @@ } }, { - "accuracy": 0.9804006579675173, + "accuracy": 0.9804631314779583, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -71666,7 +71666,7 @@ } }, { - "accuracy": 0.9826104264510306, + "accuracy": 0.9826670336095911, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -71715,7 +71715,7 @@ } }, { - "accuracy": 0.989037601571334, + "accuracy": 0.989072353432053, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -71758,7 +71758,7 @@ } }, { - "accuracy": 0.9899757441721464, + "accuracy": 0.9900079542084744, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -71801,7 +71801,7 @@ } }, { - "accuracy": 0.9891525961850819, + "accuracy": 0.9891899533961949, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -71853,7 +71853,7 @@ } }, { - "accuracy": 0.9905065509833788, + "accuracy": 0.9905387523927187, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -71905,7 +71905,7 @@ } }, { - "accuracy": 0.9944709221783438, + "accuracy": 0.9944872656150868, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -71957,7 +71957,7 @@ } }, { - "accuracy": 0.9952695142281683, + "accuracy": 0.9952858039423039, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -72009,7 +72009,7 @@ } }, { - "accuracy": 0.9970245618177088, + "accuracy": 0.9970333137010273, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -72052,7 +72052,7 @@ } }, { - "accuracy": 0.9971665218472481, + "accuracy": 0.9971760927062285, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -72101,7 +72101,7 @@ } }, { - "accuracy": 0.997586926347331, + "accuracy": 0.9975953058976876, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -72147,7 +72147,7 @@ } }, { - "accuracy": 0.9991963135176584, + "accuracy": 0.9991984134049792, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -72189,7 +72189,7 @@ ], "model.layers.37.self_attn": [ { - "accuracy": 0.9860341203840155, + "accuracy": 0.986051829237687, "total_bits": 320921600, "q_proj": { "group_size": { @@ -72253,7 +72253,7 @@ } }, { - "accuracy": 0.986598014831543, + "accuracy": 0.9866202489325875, "total_bits": 329244672, "q_proj": { "group_size": { @@ -72317,7 +72317,7 @@ } }, { - "accuracy": 0.988054160224764, + "accuracy": 0.9880674528448206, "total_bits": 336188416, "q_proj": { "group_size": { @@ -72381,7 +72381,7 @@ } }, { - "accuracy": 0.9907922799649992, + "accuracy": 0.9907888448552081, "total_bits": 401721344, "q_proj": { "group_size": { @@ -72445,7 +72445,7 @@ } }, { - "accuracy": 0.9930573976353595, + "accuracy": 0.9930680041250429, "total_bits": 475443200, "q_proj": { "group_size": { @@ -72509,7 +72509,7 @@ } }, { - "accuracy": 0.9932315106454649, + "accuracy": 0.9932365739031842, "total_bits": 475642880, "q_proj": { "group_size": { @@ -72573,7 +72573,7 @@ } }, { - "accuracy": 0.9956334288183012, + "accuracy": 0.9956420240433592, "total_bits": 609923072, "q_proj": { "group_size": { @@ -72625,7 +72625,7 @@ } }, { - "accuracy": 0.9958631588440192, + "accuracy": 0.9958811829748907, "total_bits": 610188288, "q_proj": { "group_size": { @@ -72677,7 +72677,7 @@ } }, { - "accuracy": 0.9962335907314953, + "accuracy": 0.9962233987293745, "total_bits": 615184384, "q_proj": { "group_size": { @@ -72729,7 +72729,7 @@ } }, { - "accuracy": 0.9964758851810506, + "accuracy": 0.9964635003554193, "total_bits": 624115712, "q_proj": { "group_size": { @@ -72781,7 +72781,7 @@ } }, { - "accuracy": 0.9965564188988585, + "accuracy": 0.996561628423239, "total_bits": 626637824, "q_proj": { "group_size": { @@ -72845,7 +72845,7 @@ } }, { - "accuracy": 0.9967955090105534, + "accuracy": 0.9968100390151927, "total_bits": 630519808, "q_proj": { "group_size": { @@ -72909,7 +72909,7 @@ } }, { - "accuracy": 0.9970262723141595, + "accuracy": 0.997037233490693, "total_bits": 637526016, "q_proj": { "group_size": { @@ -72970,7 +72970,7 @@ } }, { - "accuracy": 0.9972823411226273, + "accuracy": 0.9972880117987332, "total_bits": 646987776, "q_proj": { "group_size": { @@ -73031,7 +73031,7 @@ } }, { - "accuracy": 0.9983461656068501, + "accuracy": 0.9983498885443336, "total_bits": 784904192, "q_proj": { "group_size": { @@ -73092,7 +73092,7 @@ } }, { - "accuracy": 0.9986111309967542, + "accuracy": 0.9986202342337683, "total_bits": 797982720, "q_proj": { "group_size": { @@ -73153,7 +73153,7 @@ } }, { - "accuracy": 0.9988535253429099, + "accuracy": 0.9988574053307897, "total_bits": 911912960, "q_proj": { "group_size": { @@ -73205,7 +73205,7 @@ } }, { - "accuracy": 0.9992160088333645, + "accuracy": 0.9992187461374622, "total_bits": 942882816, "q_proj": { "group_size": { @@ -73257,7 +73257,7 @@ } }, { - "accuracy": 0.9996473832595113, + "accuracy": 0.9996476976053887, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -73311,7 +73311,7 @@ ], "model.layers.37.mlp": [ { - "accuracy": 0.9562807867401525, + "accuracy": 0.9564260087515178, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -73363,7 +73363,7 @@ } }, { - "accuracy": 0.95756039807671, + "accuracy": 0.957688780207383, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -73415,7 +73415,7 @@ } }, { - "accuracy": 0.9631223678588867, + "accuracy": 0.963229913460581, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -73464,7 +73464,7 @@ } }, { - "accuracy": 0.96468837951359, + "accuracy": 0.9647946357727051, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -73513,7 +73513,7 @@ } }, { - "accuracy": 0.9781959982294786, + "accuracy": 0.9782538006180211, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -73565,7 +73565,7 @@ } }, { - "accuracy": 0.9799491684687766, + "accuracy": 0.9800077959110862, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -73617,7 +73617,7 @@ } }, { - "accuracy": 0.9821940942814475, + "accuracy": 0.982245900129017, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -73666,7 +73666,7 @@ } }, { - "accuracy": 0.988819577976277, + "accuracy": 0.9888535800733065, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -73709,7 +73709,7 @@ } }, { - "accuracy": 0.9897678486610714, + "accuracy": 0.9897995512736472, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -73752,7 +73752,7 @@ } }, { - "accuracy": 0.9889270718160429, + "accuracy": 0.9889593414570156, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -73804,7 +73804,7 @@ } }, { - "accuracy": 0.9903033940415633, + "accuracy": 0.9903299502636257, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -73856,7 +73856,7 @@ } }, { - "accuracy": 0.9943593913003018, + "accuracy": 0.9943788561381792, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -73908,7 +73908,7 @@ } }, { - "accuracy": 0.9951676659678158, + "accuracy": 0.9951822208730798, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -73960,7 +73960,7 @@ } }, { - "accuracy": 0.9969648561979595, + "accuracy": 0.9969748778170661, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -74003,7 +74003,7 @@ } }, { - "accuracy": 0.9971152498925987, + "accuracy": 0.9971240803991493, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -74052,7 +74052,7 @@ } }, { - "accuracy": 0.9975433986830083, + "accuracy": 0.997551116896303, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -74098,7 +74098,7 @@ } }, { - "accuracy": 0.9991783913911173, + "accuracy": 0.9991814175032472, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -74140,7 +74140,7 @@ ], "model.layers.38.self_attn": [ { - "accuracy": 0.9832555212472615, + "accuracy": 0.9833002341421027, "total_bits": 320921600, "q_proj": { "group_size": { @@ -74204,7 +74204,7 @@ } }, { - "accuracy": 0.9839830476986734, + "accuracy": 0.9840111277605358, "total_bits": 329244672, "q_proj": { "group_size": { @@ -74268,7 +74268,7 @@ } }, { - "accuracy": 0.9855602650265944, + "accuracy": 0.9855831563472748, "total_bits": 336188416, "q_proj": { "group_size": { @@ -74332,7 +74332,7 @@ } }, { - "accuracy": 0.989836816724978, + "accuracy": 0.9898528597856823, "total_bits": 401721344, "q_proj": { "group_size": { @@ -74396,7 +74396,7 @@ } }, { - "accuracy": 0.9917040765285492, + "accuracy": 0.9917241264330713, "total_bits": 475443200, "q_proj": { "group_size": { @@ -74460,7 +74460,7 @@ } }, { - "accuracy": 0.9918553139034071, + "accuracy": 0.991868833177968, "total_bits": 475642880, "q_proj": { "group_size": { @@ -74524,7 +74524,7 @@ } }, { - "accuracy": 0.9950723071631632, + "accuracy": 0.9950934660277868, "total_bits": 609923072, "q_proj": { "group_size": { @@ -74576,7 +74576,7 @@ } }, { - "accuracy": 0.9952876160019323, + "accuracy": 0.9952884214489084, "total_bits": 610188288, "q_proj": { "group_size": { @@ -74628,7 +74628,7 @@ } }, { - "accuracy": 0.9956092116864104, + "accuracy": 0.9956187511745253, "total_bits": 615184384, "q_proj": { "group_size": { @@ -74680,7 +74680,7 @@ } }, { - "accuracy": 0.9958530323285806, + "accuracy": 0.9958718568086624, "total_bits": 624115712, "q_proj": { "group_size": { @@ -74732,7 +74732,7 @@ } }, { - "accuracy": 0.9958575430669283, + "accuracy": 0.9958644438731042, "total_bits": 626637824, "q_proj": { "group_size": { @@ -74796,7 +74796,7 @@ } }, { - "accuracy": 0.9961224904185847, + "accuracy": 0.9961309080061159, "total_bits": 630519808, "q_proj": { "group_size": { @@ -74860,7 +74860,7 @@ } }, { - "accuracy": 0.9964248922310377, + "accuracy": 0.9964320667480168, "total_bits": 637526016, "q_proj": { "group_size": { @@ -74921,7 +74921,7 @@ } }, { - "accuracy": 0.9967100643797925, + "accuracy": 0.9967168198996469, "total_bits": 646987776, "q_proj": { "group_size": { @@ -74982,7 +74982,7 @@ } }, { - "accuracy": 0.9980359118627874, + "accuracy": 0.9980394699071583, "total_bits": 784904192, "q_proj": { "group_size": { @@ -75043,7 +75043,7 @@ } }, { - "accuracy": 0.9983396936011942, + "accuracy": 0.9983434435960493, "total_bits": 797982720, "q_proj": { "group_size": { @@ -75104,7 +75104,7 @@ } }, { - "accuracy": 0.998712934945759, + "accuracy": 0.9987181016293011, "total_bits": 911912960, "q_proj": { "group_size": { @@ -75156,7 +75156,7 @@ } }, { - "accuracy": 0.9990848468518571, + "accuracy": 0.9990868316473145, "total_bits": 942882816, "q_proj": { "group_size": { @@ -75208,7 +75208,7 @@ } }, { - "accuracy": 0.9996244190750938, + "accuracy": 0.9996249874024407, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -75262,7 +75262,7 @@ ], "model.layers.38.mlp": [ { - "accuracy": 0.9559017200219004, + "accuracy": 0.9560212932134929, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -75314,7 +75314,7 @@ } }, { - "accuracy": 0.9571980583040338, + "accuracy": 0.9573194196349696, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -75366,7 +75366,7 @@ } }, { - "accuracy": 0.9629721139606676, + "accuracy": 0.9630706498497411, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -75415,7 +75415,7 @@ } }, { - "accuracy": 0.964581947577627, + "accuracy": 0.9646774122589513, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -75464,7 +75464,7 @@ } }, { - "accuracy": 0.9780515683324713, + "accuracy": 0.9781081362774497, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -75516,7 +75516,7 @@ } }, { - "accuracy": 0.9798042695773276, + "accuracy": 0.9798582149179358, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -75568,7 +75568,7 @@ } }, { - "accuracy": 0.982119525733747, + "accuracy": 0.9821650872105047, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -75617,7 +75617,7 @@ } }, { - "accuracy": 0.9887246183658901, + "accuracy": 0.9887553908322987, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -75660,7 +75660,7 @@ } }, { - "accuracy": 0.9896894881599828, + "accuracy": 0.9897180356477436, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -75703,7 +75703,7 @@ } }, { - "accuracy": 0.9888435570817244, + "accuracy": 0.9888723320082614, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -75755,7 +75755,7 @@ } }, { - "accuracy": 0.990226467189036, + "accuracy": 0.9902532139891073, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -75807,7 +75807,7 @@ } }, { - "accuracy": 0.9943145317466635, + "accuracy": 0.9943290168517515, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -75859,7 +75859,7 @@ } }, { - "accuracy": 0.9951314459505834, + "accuracy": 0.9951424684963728, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -75911,7 +75911,7 @@ } }, { - "accuracy": 0.9969406525947546, + "accuracy": 0.9969502465897485, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -75954,7 +75954,7 @@ } }, { - "accuracy": 0.9970917240961602, + "accuracy": 0.9970994432897944, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -76003,7 +76003,7 @@ } }, { - "accuracy": 0.9975315114777339, + "accuracy": 0.9975378148649868, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -76049,7 +76049,7 @@ } }, { - "accuracy": 0.9991807503331649, + "accuracy": 0.9991826381240236, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -76091,7 +76091,7 @@ ], "model.layers.39.self_attn": [ { - "accuracy": 0.9855250731894845, + "accuracy": 0.9855525148542303, "total_bits": 320921600, "q_proj": { "group_size": { @@ -76155,7 +76155,7 @@ } }, { - "accuracy": 0.9861719200485631, + "accuracy": 0.986172710594378, "total_bits": 329244672, "q_proj": { "group_size": { @@ -76219,7 +76219,7 @@ } }, { - "accuracy": 0.9876591720079121, + "accuracy": 0.987674283353906, "total_bits": 336188416, "q_proj": { "group_size": { @@ -76283,7 +76283,7 @@ } }, { - "accuracy": 0.9909688030418596, + "accuracy": 0.9909755156228417, "total_bits": 401721344, "q_proj": { "group_size": { @@ -76347,7 +76347,7 @@ } }, { - "accuracy": 0.9927946956534135, + "accuracy": 0.9928264108143354, "total_bits": 475443200, "q_proj": { "group_size": { @@ -76411,7 +76411,7 @@ } }, { - "accuracy": 0.9929800457076022, + "accuracy": 0.9929954017463484, "total_bits": 475642880, "q_proj": { "group_size": { @@ -76475,7 +76475,7 @@ } }, { - "accuracy": 0.9956030088819956, + "accuracy": 0.9956150647056731, "total_bits": 609923072, "q_proj": { "group_size": { @@ -76527,7 +76527,7 @@ } }, { - "accuracy": 0.9958557984546611, + "accuracy": 0.9958553863199133, "total_bits": 610188288, "q_proj": { "group_size": { @@ -76579,7 +76579,7 @@ } }, { - "accuracy": 0.9961979953866256, + "accuracy": 0.9962042425023881, "total_bits": 615184384, "q_proj": { "group_size": { @@ -76631,7 +76631,7 @@ } }, { - "accuracy": 0.9964241103122109, + "accuracy": 0.9964296170755437, "total_bits": 624115712, "q_proj": { "group_size": { @@ -76683,7 +76683,7 @@ } }, { - "accuracy": 0.9964286728124869, + "accuracy": 0.9964444378488942, "total_bits": 626637824, "q_proj": { "group_size": { @@ -76747,7 +76747,7 @@ } }, { - "accuracy": 0.9966781433475645, + "accuracy": 0.9966831226882181, "total_bits": 630519808, "q_proj": { "group_size": { @@ -76811,7 +76811,7 @@ } }, { - "accuracy": 0.996907544175261, + "accuracy": 0.9969155711955145, "total_bits": 637526016, "q_proj": { "group_size": { @@ -76872,7 +76872,7 @@ } }, { - "accuracy": 0.9971790690171091, + "accuracy": 0.9971880218700359, "total_bits": 646987776, "q_proj": { "group_size": { @@ -76933,7 +76933,7 @@ } }, { - "accuracy": 0.9983011821382924, + "accuracy": 0.9983091230847334, "total_bits": 784904192, "q_proj": { "group_size": { @@ -76994,7 +76994,7 @@ } }, { - "accuracy": 0.9985807728218404, + "accuracy": 0.9985842909663916, "total_bits": 797982720, "q_proj": { "group_size": { @@ -77055,7 +77055,7 @@ } }, { - "accuracy": 0.9988478768224779, + "accuracy": 0.9988537046470141, "total_bits": 911912960, "q_proj": { "group_size": { @@ -77107,7 +77107,7 @@ } }, { - "accuracy": 0.9992132794210001, + "accuracy": 0.9992130834511236, "total_bits": 942882816, "q_proj": { "group_size": { @@ -77159,7 +77159,7 @@ } }, { - "accuracy": 0.9996519149036, + "accuracy": 0.9996526309436089, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -77213,7 +77213,7 @@ ], "model.layers.39.mlp": [ { - "accuracy": 0.9553051057614779, + "accuracy": 0.9554182353772616, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -77265,7 +77265,7 @@ } }, { - "accuracy": 0.9566003837083515, + "accuracy": 0.9567066274191204, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -77317,7 +77317,7 @@ } }, { - "accuracy": 0.9624542280247337, + "accuracy": 0.9625533819198608, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -77366,7 +77366,7 @@ } }, { - "accuracy": 0.9640905637490123, + "accuracy": 0.9641835062127364, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -77415,7 +77415,7 @@ } }, { - "accuracy": 0.9777595385124809, + "accuracy": 0.9778230849065279, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -77467,7 +77467,7 @@ } }, { - "accuracy": 0.9795238218809429, + "accuracy": 0.979579884754984, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -77519,7 +77519,7 @@ } }, { - "accuracy": 0.9818844026640842, + "accuracy": 0.9819318369815224, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -77568,7 +77568,7 @@ } }, { - "accuracy": 0.9886049981179991, + "accuracy": 0.988638773560524, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -77611,7 +77611,7 @@ } }, { - "accuracy": 0.9895774855425483, + "accuracy": 0.9896050566121152, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -77654,7 +77654,7 @@ } }, { - "accuracy": 0.9887190610170364, + "accuracy": 0.9887515086876718, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -77706,7 +77706,7 @@ } }, { - "accuracy": 0.9901103142060732, + "accuracy": 0.9901354328582161, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -77758,7 +77758,7 @@ } }, { - "accuracy": 0.9942646116802567, + "accuracy": 0.9942824020981789, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -77810,7 +77810,7 @@ } }, { - "accuracy": 0.9950796750031019, + "accuracy": 0.9950920990423152, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -77862,7 +77862,7 @@ } }, { - "accuracy": 0.9969274244810405, + "accuracy": 0.9969365320315486, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -77905,7 +77905,7 @@ } }, { - "accuracy": 0.9970674938277194, + "accuracy": 0.997075582413297, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -77954,7 +77954,7 @@ } }, { - "accuracy": 0.997512735230358, + "accuracy": 0.9975199675873706, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -78000,7 +78000,7 @@ } }, { - "accuracy": 0.9991773777690373, + "accuracy": 0.9991801554631246, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -78042,7 +78042,7 @@ ], "model.layers.40.self_attn": [ { - "accuracy": 0.9814184778615048, + "accuracy": 0.9814315124561912, "total_bits": 320921600, "q_proj": { "group_size": { @@ -78106,7 +78106,7 @@ } }, { - "accuracy": 0.9821068911175979, + "accuracy": 0.982159845138851, "total_bits": 329244672, "q_proj": { "group_size": { @@ -78170,7 +78170,7 @@ } }, { - "accuracy": 0.9837647331388373, + "accuracy": 0.9837938045200548, "total_bits": 336188416, "q_proj": { "group_size": { @@ -78234,7 +78234,7 @@ } }, { - "accuracy": 0.9883846068068555, + "accuracy": 0.9884339743538907, "total_bits": 401721344, "q_proj": { "group_size": { @@ -78298,7 +78298,7 @@ } }, { - "accuracy": 0.9905623001487631, + "accuracy": 0.990576435076563, "total_bits": 475443200, "q_proj": { "group_size": { @@ -78362,7 +78362,7 @@ } }, { - "accuracy": 0.9908519253918999, + "accuracy": 0.9908744551633534, "total_bits": 475642880, "q_proj": { "group_size": { @@ -78426,7 +78426,7 @@ } }, { - "accuracy": 0.9940167975268865, + "accuracy": 0.9940018618577405, "total_bits": 609923072, "q_proj": { "group_size": { @@ -78478,7 +78478,7 @@ } }, { - "accuracy": 0.9943952991774208, + "accuracy": 0.9944093807747489, "total_bits": 610188288, "q_proj": { "group_size": { @@ -78530,7 +78530,7 @@ } }, { - "accuracy": 0.9949455767085678, + "accuracy": 0.9949546338696229, "total_bits": 615184384, "q_proj": { "group_size": { @@ -78582,7 +78582,7 @@ } }, { - "accuracy": 0.9952138011392794, + "accuracy": 0.9952287693556986, "total_bits": 624115712, "q_proj": { "group_size": { @@ -78634,7 +78634,7 @@ } }, { - "accuracy": 0.9953307156500063, + "accuracy": 0.9953440039565689, "total_bits": 626637824, "q_proj": { "group_size": { @@ -78698,7 +78698,7 @@ } }, { - "accuracy": 0.9956705852558738, + "accuracy": 0.9956802608151185, "total_bits": 630519808, "q_proj": { "group_size": { @@ -78762,7 +78762,7 @@ } }, { - "accuracy": 0.9959101967121425, + "accuracy": 0.9959225180117708, "total_bits": 637526016, "q_proj": { "group_size": { @@ -78823,7 +78823,7 @@ } }, { - "accuracy": 0.9962732662495813, + "accuracy": 0.9962762429525978, "total_bits": 646987776, "q_proj": { "group_size": { @@ -78884,7 +78884,7 @@ } }, { - "accuracy": 0.9977015356876349, + "accuracy": 0.9977109730243683, "total_bits": 784904192, "q_proj": { "group_size": { @@ -78945,7 +78945,7 @@ } }, { - "accuracy": 0.9981104164923492, + "accuracy": 0.9981103225758201, "total_bits": 797982720, "q_proj": { "group_size": { @@ -79006,7 +79006,7 @@ } }, { - "accuracy": 0.9984334968030453, + "accuracy": 0.9984395427531317, "total_bits": 911912960, "q_proj": { "group_size": { @@ -79058,7 +79058,7 @@ } }, { - "accuracy": 0.9989452987517181, + "accuracy": 0.9989462234079838, "total_bits": 942882816, "q_proj": { "group_size": { @@ -79110,7 +79110,7 @@ } }, { - "accuracy": 0.9995370752815353, + "accuracy": 0.9995383627124523, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -79164,7 +79164,7 @@ ], "model.layers.40.mlp": [ { - "accuracy": 0.9540786805905794, + "accuracy": 0.9541957284274855, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -79216,7 +79216,7 @@ } }, { - "accuracy": 0.9553763929166292, + "accuracy": 0.9554943912907651, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -79268,7 +79268,7 @@ } }, { - "accuracy": 0.9613418108538577, + "accuracy": 0.9614489894164235, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -79317,7 +79317,7 @@ } }, { - "accuracy": 0.9630169303793656, + "accuracy": 0.9631223208025882, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -79366,7 +79366,7 @@ } }, { - "accuracy": 0.9771259458441484, + "accuracy": 0.9771755133804522, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -79418,7 +79418,7 @@ } }, { - "accuracy": 0.9789606677858453, + "accuracy": 0.9790080691638746, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -79470,7 +79470,7 @@ } }, { - "accuracy": 0.9813736928136725, + "accuracy": 0.9814160309339824, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -79519,7 +79519,7 @@ } }, { - "accuracy": 0.9882696325841703, + "accuracy": 0.9882933818980267, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -79562,7 +79562,7 @@ } }, { - "accuracy": 0.9892762538633848, + "accuracy": 0.989298939704895, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -79605,7 +79605,7 @@ } }, { - "accuracy": 0.98838451818416, + "accuracy": 0.9884130201841655, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -79657,7 +79657,7 @@ } }, { - "accuracy": 0.9898295426055005, + "accuracy": 0.9898540769752703, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -79709,7 +79709,7 @@ } }, { - "accuracy": 0.9940878748893738, + "accuracy": 0.9941009714415199, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -79761,7 +79761,7 @@ } }, { - "accuracy": 0.9949347208205023, + "accuracy": 0.9949471213315663, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -79813,7 +79813,7 @@ } }, { - "accuracy": 0.9968236620096784, + "accuracy": 0.9968316643253753, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -79856,7 +79856,7 @@ } }, { - "accuracy": 0.9969757014983579, + "accuracy": 0.9969830460062152, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -79905,7 +79905,7 @@ } }, { - "accuracy": 0.9974281978058187, + "accuracy": 0.997435115669903, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -79951,7 +79951,7 @@ } }, { - "accuracy": 0.9991336228620065, + "accuracy": 0.9991359013671938, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -79993,7 +79993,7 @@ ], "model.layers.41.self_attn": [ { - "accuracy": 0.9802216338483911, + "accuracy": 0.9802678996010831, "total_bits": 320921600, "q_proj": { "group_size": { @@ -80057,7 +80057,7 @@ } }, { - "accuracy": 0.9811765912332033, + "accuracy": 0.9811990841438896, "total_bits": 329244672, "q_proj": { "group_size": { @@ -80121,7 +80121,7 @@ } }, { - "accuracy": 0.9828440327393381, + "accuracy": 0.9828641195046274, "total_bits": 336188416, "q_proj": { "group_size": { @@ -80185,7 +80185,7 @@ } }, { - "accuracy": 0.9884208683904848, + "accuracy": 0.9884371286944339, "total_bits": 401721344, "q_proj": { "group_size": { @@ -80249,7 +80249,7 @@ } }, { - "accuracy": 0.9901486180330578, + "accuracy": 0.9901655896713859, "total_bits": 475443200, "q_proj": { "group_size": { @@ -80313,7 +80313,7 @@ } }, { - "accuracy": 0.990317873264614, + "accuracy": 0.9903380729650196, "total_bits": 475642880, "q_proj": { "group_size": { @@ -80377,7 +80377,7 @@ } }, { - "accuracy": 0.9940311665597715, + "accuracy": 0.9940469021859922, "total_bits": 609923072, "q_proj": { "group_size": { @@ -80429,7 +80429,7 @@ } }, { - "accuracy": 0.9942567615132583, + "accuracy": 0.9942707450766313, "total_bits": 610188288, "q_proj": { "group_size": { @@ -80481,7 +80481,7 @@ } }, { - "accuracy": 0.9946994326616588, + "accuracy": 0.9947012784449678, "total_bits": 615184384, "q_proj": { "group_size": { @@ -80533,7 +80533,7 @@ } }, { - "accuracy": 0.9950137350120043, + "accuracy": 0.9950247010902354, "total_bits": 624115712, "q_proj": { "group_size": { @@ -80585,7 +80585,7 @@ } }, { - "accuracy": 0.9950724448028364, + "accuracy": 0.9950795318735274, "total_bits": 626637824, "q_proj": { "group_size": { @@ -80649,7 +80649,7 @@ } }, { - "accuracy": 0.9953864138377341, + "accuracy": 0.9953869487109938, "total_bits": 630519808, "q_proj": { "group_size": { @@ -80713,7 +80713,7 @@ } }, { - "accuracy": 0.9956942267323795, + "accuracy": 0.9956950506097392, "total_bits": 637526016, "q_proj": { "group_size": { @@ -80774,7 +80774,7 @@ } }, { - "accuracy": 0.9960487586887259, + "accuracy": 0.9960519083236393, "total_bits": 646987776, "q_proj": { "group_size": { @@ -80835,7 +80835,7 @@ } }, { - "accuracy": 0.9976414316578915, + "accuracy": 0.9976446922672423, "total_bits": 784904192, "q_proj": { "group_size": { @@ -80896,7 +80896,7 @@ } }, { - "accuracy": 0.9980102740228176, + "accuracy": 0.9980148133870802, "total_bits": 797982720, "q_proj": { "group_size": { @@ -80957,7 +80957,7 @@ } }, { - "accuracy": 0.9984498740615029, + "accuracy": 0.9984512976125667, "total_bits": 911912960, "q_proj": { "group_size": { @@ -81009,7 +81009,7 @@ } }, { - "accuracy": 0.9989002655799452, + "accuracy": 0.9989009447591869, "total_bits": 942882816, "q_proj": { "group_size": { @@ -81061,7 +81061,7 @@ } }, { - "accuracy": 0.9995545468440181, + "accuracy": 0.999555171222279, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -81115,7 +81115,7 @@ ], "model.layers.41.mlp": [ { - "accuracy": 0.9524776966948258, + "accuracy": 0.9525854650296663, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -81167,7 +81167,7 @@ } }, { - "accuracy": 0.9538698290523729, + "accuracy": 0.953967295194927, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -81219,7 +81219,7 @@ } }, { - "accuracy": 0.9603133860387301, + "accuracy": 0.9603997186610573, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -81268,7 +81268,7 @@ } }, { - "accuracy": 0.9620985608351857, + "accuracy": 0.9621780012783251, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -81317,7 +81317,7 @@ } }, { - "accuracy": 0.9764034653964796, + "accuracy": 0.97645587356467, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -81369,7 +81369,7 @@ } }, { - "accuracy": 0.9782825316253462, + "accuracy": 0.9783347954875544, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -81421,7 +81421,7 @@ } }, { - "accuracy": 0.9808455297821447, + "accuracy": 0.9808924856938814, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -81470,7 +81470,7 @@ } }, { - "accuracy": 0.9879061995368255, + "accuracy": 0.987930217855855, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -81513,7 +81513,7 @@ } }, { - "accuracy": 0.9889350674654308, + "accuracy": 0.9889622816914007, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -81556,7 +81556,7 @@ } }, { - "accuracy": 0.9880294078274777, + "accuracy": 0.9880516450656088, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -81608,7 +81608,7 @@ } }, { - "accuracy": 0.989509188815167, + "accuracy": 0.9895327122587907, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -81660,7 +81660,7 @@ } }, { - "accuracy": 0.99391208589077, + "accuracy": 0.9939233150921369, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -81712,7 +81712,7 @@ } }, { - "accuracy": 0.9947800048087773, + "accuracy": 0.9947911042131876, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -81764,7 +81764,7 @@ } }, { - "accuracy": 0.9967347197234631, + "accuracy": 0.9967404354951883, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -81807,7 +81807,7 @@ } }, { - "accuracy": 0.9968879216987836, + "accuracy": 0.9968932263161007, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -81856,7 +81856,7 @@ } }, { - "accuracy": 0.9973754453423777, + "accuracy": 0.9973796870755521, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -81902,7 +81902,7 @@ } }, { - "accuracy": 0.9991298015964659, + "accuracy": 0.9991314018047169, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -81944,7 +81944,7 @@ ], "model.layers.42.self_attn": [ { - "accuracy": 0.9821137425146604, + "accuracy": 0.9821364220819975, "total_bits": 320921600, "q_proj": { "group_size": { @@ -82008,7 +82008,7 @@ } }, { - "accuracy": 0.9828916129313017, + "accuracy": 0.9829194796712775, "total_bits": 329244672, "q_proj": { "group_size": { @@ -82072,7 +82072,7 @@ } }, { - "accuracy": 0.9846285079654894, + "accuracy": 0.9846489711811668, "total_bits": 336188416, "q_proj": { "group_size": { @@ -82136,7 +82136,7 @@ } }, { - "accuracy": 0.9889818414261466, + "accuracy": 0.9890014952734897, "total_bits": 401721344, "q_proj": { "group_size": { @@ -82200,7 +82200,7 @@ } }, { - "accuracy": 0.9911224528362876, + "accuracy": 0.9911259428450936, "total_bits": 475443200, "q_proj": { "group_size": { @@ -82264,7 +82264,7 @@ } }, { - "accuracy": 0.991245058806319, + "accuracy": 0.9912501463764593, "total_bits": 475642880, "q_proj": { "group_size": { @@ -82328,7 +82328,7 @@ } }, { - "accuracy": 0.9947148984984348, + "accuracy": 0.9947119390493945, "total_bits": 609923072, "q_proj": { "group_size": { @@ -82380,7 +82380,7 @@ } }, { - "accuracy": 0.9948560913142405, + "accuracy": 0.9948554384080988, "total_bits": 610188288, "q_proj": { "group_size": { @@ -82432,7 +82432,7 @@ } }, { - "accuracy": 0.9952116843901182, + "accuracy": 0.9952143007203152, "total_bits": 615184384, "q_proj": { "group_size": { @@ -82484,7 +82484,7 @@ } }, { - "accuracy": 0.9954693262514315, + "accuracy": 0.9954772646489897, "total_bits": 624115712, "q_proj": { "group_size": { @@ -82536,7 +82536,7 @@ } }, { - "accuracy": 0.9955504881708246, + "accuracy": 0.9955460507618753, "total_bits": 626637824, "q_proj": { "group_size": { @@ -82600,7 +82600,7 @@ } }, { - "accuracy": 0.9958256028200451, + "accuracy": 0.9958209932634705, "total_bits": 630519808, "q_proj": { "group_size": { @@ -82664,7 +82664,7 @@ } }, { - "accuracy": 0.9961590978660082, + "accuracy": 0.9961589927736082, "total_bits": 637526016, "q_proj": { "group_size": { @@ -82725,7 +82725,7 @@ } }, { - "accuracy": 0.9964609624523866, + "accuracy": 0.9964639211171552, "total_bits": 646987776, "q_proj": { "group_size": { @@ -82786,7 +82786,7 @@ } }, { - "accuracy": 0.9978925328897802, + "accuracy": 0.9978927650341862, "total_bits": 784904192, "q_proj": { "group_size": { @@ -82847,7 +82847,7 @@ } }, { - "accuracy": 0.9982153269413271, + "accuracy": 0.9982180401290718, "total_bits": 797982720, "q_proj": { "group_size": { @@ -82908,7 +82908,7 @@ } }, { - "accuracy": 0.9986006767538033, + "accuracy": 0.9986003819657, "total_bits": 911912960, "q_proj": { "group_size": { @@ -82960,7 +82960,7 @@ } }, { - "accuracy": 0.9990040370704312, + "accuracy": 0.9990043896005342, "total_bits": 942882816, "q_proj": { "group_size": { @@ -83012,7 +83012,7 @@ } }, { - "accuracy": 0.999585837370863, + "accuracy": 0.9995864796893377, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -83066,7 +83066,7 @@ ], "model.layers.42.mlp": [ { - "accuracy": 0.950714930107719, + "accuracy": 0.9508256629893654, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -83118,7 +83118,7 @@ } }, { - "accuracy": 0.9521572119311282, + "accuracy": 0.9522735382381239, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -83170,7 +83170,7 @@ } }, { - "accuracy": 0.9589675445305674, + "accuracy": 0.9590725553663153, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -83219,7 +83219,7 @@ } }, { - "accuracy": 0.9608522465354518, + "accuracy": 0.9609525454671759, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -83268,7 +83268,7 @@ } }, { - "accuracy": 0.9755896925926208, + "accuracy": 0.9756484784578022, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -83320,7 +83320,7 @@ } }, { - "accuracy": 0.9775337012190568, + "accuracy": 0.9775775357296592, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -83372,7 +83372,7 @@ } }, { - "accuracy": 0.9802326607076746, + "accuracy": 0.980273027169077, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -83421,7 +83421,7 @@ } }, { - "accuracy": 0.9874629793982757, + "accuracy": 0.9874963823117708, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -83464,7 +83464,7 @@ } }, { - "accuracy": 0.9885397286791551, + "accuracy": 0.9885658261023069, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -83507,7 +83507,7 @@ } }, { - "accuracy": 0.9876113659457156, + "accuracy": 0.9876434638311988, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -83559,7 +83559,7 @@ } }, { - "accuracy": 0.989148420722861, + "accuracy": 0.9891741024820428, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -83611,7 +83611,7 @@ } }, { - "accuracy": 0.9936986796950039, + "accuracy": 0.9937167367652843, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -83663,7 +83663,7 @@ } }, { - "accuracy": 0.9945994097935525, + "accuracy": 0.994611279744851, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -83715,7 +83715,7 @@ } }, { - "accuracy": 0.9966152667214996, + "accuracy": 0.9966245046571681, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -83758,7 +83758,7 @@ } }, { - "accuracy": 0.9967810219447863, + "accuracy": 0.996789673049199, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -83807,7 +83807,7 @@ } }, { - "accuracy": 0.9973012638327322, + "accuracy": 0.9973091628206404, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -83853,7 +83853,7 @@ } }, { - "accuracy": 0.9990991938271021, + "accuracy": 0.9991012950868983, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -83895,7 +83895,7 @@ ], "model.layers.43.self_attn": [ { - "accuracy": 0.9790838784293124, + "accuracy": 0.9791207784100583, "total_bits": 320921600, "q_proj": { "group_size": { @@ -83959,7 +83959,7 @@ } }, { - "accuracy": 0.9799497692208541, + "accuracy": 0.9800180024222324, "total_bits": 329244672, "q_proj": { "group_size": { @@ -84023,7 +84023,7 @@ } }, { - "accuracy": 0.9815491688878912, + "accuracy": 0.9816259710412276, "total_bits": 336188416, "q_proj": { "group_size": { @@ -84087,7 +84087,7 @@ } }, { - "accuracy": 0.9868392928650505, + "accuracy": 0.9868980097143274, "total_bits": 401721344, "q_proj": { "group_size": { @@ -84151,7 +84151,7 @@ } }, { - "accuracy": 0.9894877334958628, + "accuracy": 0.9895258922325937, "total_bits": 475443200, "q_proj": { "group_size": { @@ -84215,7 +84215,7 @@ } }, { - "accuracy": 0.989665886289195, + "accuracy": 0.9896818964104903, "total_bits": 475642880, "q_proj": { "group_size": { @@ -84279,7 +84279,7 @@ } }, { - "accuracy": 0.9934717095211932, + "accuracy": 0.9934989817832646, "total_bits": 609923072, "q_proj": { "group_size": { @@ -84331,7 +84331,7 @@ } }, { - "accuracy": 0.9936990514397621, + "accuracy": 0.9937013571983889, "total_bits": 610188288, "q_proj": { "group_size": { @@ -84383,7 +84383,7 @@ } }, { - "accuracy": 0.9942434983033883, + "accuracy": 0.9942453033045718, "total_bits": 615184384, "q_proj": { "group_size": { @@ -84435,7 +84435,7 @@ } }, { - "accuracy": 0.9945849627256393, + "accuracy": 0.9945960299749124, "total_bits": 624115712, "q_proj": { "group_size": { @@ -84487,7 +84487,7 @@ } }, { - "accuracy": 0.9947096285851378, + "accuracy": 0.9947262680844257, "total_bits": 626637824, "q_proj": { "group_size": { @@ -84551,7 +84551,7 @@ } }, { - "accuracy": 0.9950776437395498, + "accuracy": 0.9950944491122898, "total_bits": 630519808, "q_proj": { "group_size": { @@ -84615,7 +84615,7 @@ } }, { - "accuracy": 0.9953358804709033, + "accuracy": 0.9953469955607465, "total_bits": 637526016, "q_proj": { "group_size": { @@ -84676,7 +84676,7 @@ } }, { - "accuracy": 0.9957301432364866, + "accuracy": 0.9957384698485073, "total_bits": 646987776, "q_proj": { "group_size": { @@ -84737,7 +84737,7 @@ } }, { - "accuracy": 0.9974198213925487, + "accuracy": 0.9974248422995994, "total_bits": 784904192, "q_proj": { "group_size": { @@ -84798,7 +84798,7 @@ } }, { - "accuracy": 0.9978369686164354, + "accuracy": 0.9978394986767518, "total_bits": 797982720, "q_proj": { "group_size": { @@ -84859,7 +84859,7 @@ } }, { - "accuracy": 0.9982898170618635, + "accuracy": 0.9982902082173448, "total_bits": 911912960, "q_proj": { "group_size": { @@ -84911,7 +84911,7 @@ } }, { - "accuracy": 0.9987713749471464, + "accuracy": 0.998776422127297, "total_bits": 942882816, "q_proj": { "group_size": { @@ -84963,7 +84963,7 @@ } }, { - "accuracy": 0.9994775397016814, + "accuracy": 0.9994786273884145, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -85017,7 +85017,7 @@ ], "model.layers.43.mlp": [ { - "accuracy": 0.9490579868617811, + "accuracy": 0.9491652250289917, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -85069,7 +85069,7 @@ } }, { - "accuracy": 0.9505483664964375, + "accuracy": 0.9506589205641496, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -85121,7 +85121,7 @@ } }, { - "accuracy": 0.9577134031998484, + "accuracy": 0.957802763110713, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -85170,7 +85170,7 @@ } }, { - "accuracy": 0.9596757230005766, + "accuracy": 0.9597640821808263, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -85219,7 +85219,7 @@ } }, { - "accuracy": 0.9747862941340396, + "accuracy": 0.9748418268404508, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -85271,7 +85271,7 @@ } }, { - "accuracy": 0.9768149335133401, + "accuracy": 0.9768629670143127, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -85323,7 +85323,7 @@ } }, { - "accuracy": 0.9796290303531446, + "accuracy": 0.9796701152073709, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -85372,7 +85372,7 @@ } }, { - "accuracy": 0.9870383974752928, + "accuracy": 0.9870658038478148, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -85415,7 +85415,7 @@ } }, { - "accuracy": 0.988154441902512, + "accuracy": 0.988180935382843, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -85458,7 +85458,7 @@ } }, { - "accuracy": 0.9871999268469057, + "accuracy": 0.9872233302969682, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -85510,7 +85510,7 @@ } }, { - "accuracy": 0.9887912743969968, + "accuracy": 0.9888185231309188, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -85562,7 +85562,7 @@ } }, { - "accuracy": 0.9934830108755514, + "accuracy": 0.9934960407646078, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -85614,7 +85614,7 @@ } }, { - "accuracy": 0.9944172556463041, + "accuracy": 0.9944301094663771, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -85666,7 +85666,7 @@ } }, { - "accuracy": 0.9964921552883951, + "accuracy": 0.9964992200073443, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -85709,7 +85709,7 @@ } }, { - "accuracy": 0.9966695179280481, + "accuracy": 0.9966762124707824, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -85758,7 +85758,7 @@ } }, { - "accuracy": 0.9972150761045908, + "accuracy": 0.9972207875628221, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -85804,7 +85804,7 @@ } }, { - "accuracy": 0.9990623156098943, + "accuracy": 0.9990643501085671, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -85846,7 +85846,7 @@ ], "model.layers.44.self_attn": [ { - "accuracy": 0.9781118662733781, + "accuracy": 0.9781102742019453, "total_bits": 320921600, "q_proj": { "group_size": { @@ -85910,7 +85910,7 @@ } }, { - "accuracy": 0.9789538477596483, + "accuracy": 0.978978759364078, "total_bits": 329244672, "q_proj": { "group_size": { @@ -85974,7 +85974,7 @@ } }, { - "accuracy": 0.9806047442712282, + "accuracy": 0.9806308071864279, "total_bits": 336188416, "q_proj": { "group_size": { @@ -86038,7 +86038,7 @@ } }, { - "accuracy": 0.9859242454955452, + "accuracy": 0.9859545826911926, "total_bits": 401721344, "q_proj": { "group_size": { @@ -86102,7 +86102,7 @@ } }, { - "accuracy": 0.9890308042890147, + "accuracy": 0.9890457046659369, "total_bits": 475443200, "q_proj": { "group_size": { @@ -86166,7 +86166,7 @@ } }, { - "accuracy": 0.9892034758078424, + "accuracy": 0.9892165951038662, "total_bits": 475642880, "q_proj": { "group_size": { @@ -86230,7 +86230,7 @@ } }, { - "accuracy": 0.9933803089355168, + "accuracy": 0.9933874756097794, "total_bits": 609923072, "q_proj": { "group_size": { @@ -86282,7 +86282,7 @@ } }, { - "accuracy": 0.9936090915610916, + "accuracy": 0.993619487865975, "total_bits": 610188288, "q_proj": { "group_size": { @@ -86334,7 +86334,7 @@ } }, { - "accuracy": 0.9941168294141167, + "accuracy": 0.994142337064994, "total_bits": 615184384, "q_proj": { "group_size": { @@ -86386,7 +86386,7 @@ } }, { - "accuracy": 0.9944906438651838, + "accuracy": 0.9945105655412925, "total_bits": 624115712, "q_proj": { "group_size": { @@ -86438,7 +86438,7 @@ } }, { - "accuracy": 0.9945107404338686, + "accuracy": 0.994524427542561, "total_bits": 626637824, "q_proj": { "group_size": { @@ -86502,7 +86502,7 @@ } }, { - "accuracy": 0.9948710316890165, + "accuracy": 0.9948754612552492, "total_bits": 630519808, "q_proj": { "group_size": { @@ -86566,7 +86566,7 @@ } }, { - "accuracy": 0.9951701838719217, + "accuracy": 0.9951642010557024, "total_bits": 637526016, "q_proj": { "group_size": { @@ -86627,7 +86627,7 @@ } }, { - "accuracy": 0.9955601064782393, + "accuracy": 0.9955632247422871, "total_bits": 646987776, "q_proj": { "group_size": { @@ -86688,7 +86688,7 @@ } }, { - "accuracy": 0.9973455373393861, + "accuracy": 0.9973488726506108, "total_bits": 784904192, "q_proj": { "group_size": { @@ -86749,7 +86749,7 @@ } }, { - "accuracy": 0.9977575848369222, + "accuracy": 0.9977557359165267, "total_bits": 797982720, "q_proj": { "group_size": { @@ -86810,7 +86810,7 @@ } }, { - "accuracy": 0.9982859504066015, + "accuracy": 0.9982909107286679, "total_bits": 911912960, "q_proj": { "group_size": { @@ -86862,7 +86862,7 @@ } }, { - "accuracy": 0.9987456625033366, + "accuracy": 0.998733692557404, "total_bits": 942882816, "q_proj": { "group_size": { @@ -86914,7 +86914,7 @@ } }, { - "accuracy": 0.9994924635577359, + "accuracy": 0.9994914016049159, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -86968,7 +86968,7 @@ ], "model.layers.44.mlp": [ { - "accuracy": 0.9461754811437506, + "accuracy": 0.9462931909059223, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -87020,7 +87020,7 @@ } }, { - "accuracy": 0.947783605048531, + "accuracy": 0.9478966750596699, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -87072,7 +87072,7 @@ } }, { - "accuracy": 0.9555485123082211, + "accuracy": 0.95563416731985, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -87121,7 +87121,7 @@ } }, { - "accuracy": 0.957670647847025, + "accuracy": 0.9577559232711792, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -87170,7 +87170,7 @@ } }, { - "accuracy": 0.973380750731418, + "accuracy": 0.9734234056974712, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -87222,7 +87222,7 @@ } }, { - "accuracy": 0.9755758956858986, + "accuracy": 0.9756327161663457, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -87274,7 +87274,7 @@ } }, { - "accuracy": 0.9786205511344107, + "accuracy": 0.9786679070246848, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -87323,7 +87323,7 @@ } }, { - "accuracy": 0.9862553367489263, + "accuracy": 0.9862792711508902, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -87366,7 +87366,7 @@ } }, { - "accuracy": 0.9874448603705356, + "accuracy": 0.9874761575146725, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -87409,7 +87409,7 @@ } }, { - "accuracy": 0.9864740261906072, + "accuracy": 0.9864957709061472, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -87461,7 +87461,7 @@ } }, { - "accuracy": 0.9881835509287683, + "accuracy": 0.9882110169059352, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -87513,7 +87513,7 @@ } }, { - "accuracy": 0.9931062914823231, + "accuracy": 0.993117534016308, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -87565,7 +87565,7 @@ } }, { - "accuracy": 0.994108929445869, + "accuracy": 0.9941216107261809, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -87617,7 +87617,7 @@ } }, { - "accuracy": 0.9962551836904726, + "accuracy": 0.9962607206482637, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -87660,7 +87660,7 @@ } }, { - "accuracy": 0.9964758432225177, + "accuracy": 0.9964809060881012, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -87709,7 +87709,7 @@ } }, { - "accuracy": 0.9970631768044672, + "accuracy": 0.997066152527144, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -87755,7 +87755,7 @@ } }, { - "accuracy": 0.9989662697832835, + "accuracy": 0.9989681583093969, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -87797,7 +87797,7 @@ ], "model.layers.45.self_attn": [ { - "accuracy": 0.9698858072883204, + "accuracy": 0.9695953538543299, "total_bits": 320921600, "q_proj": { "group_size": { @@ -87861,7 +87861,7 @@ } }, { - "accuracy": 0.9716337323188782, + "accuracy": 0.9716311379482871, "total_bits": 329244672, "q_proj": { "group_size": { @@ -87925,7 +87925,7 @@ } }, { - "accuracy": 0.9736050932030929, + "accuracy": 0.9734976762219479, "total_bits": 336188416, "q_proj": { "group_size": { @@ -87989,7 +87989,7 @@ } }, { - "accuracy": 0.9789944259743941, + "accuracy": 0.9788378775119781, "total_bits": 401721344, "q_proj": { "group_size": { @@ -88053,7 +88053,7 @@ } }, { - "accuracy": 0.9850784132355138, + "accuracy": 0.9850630415113348, "total_bits": 475443200, "q_proj": { "group_size": { @@ -88117,7 +88117,7 @@ } }, { - "accuracy": 0.985236406326294, + "accuracy": 0.9852675262250399, "total_bits": 475642880, "q_proj": { "group_size": { @@ -88181,7 +88181,7 @@ } }, { - "accuracy": 0.990776923926253, + "accuracy": 0.9907257745140478, "total_bits": 609923072, "q_proj": { "group_size": { @@ -88233,7 +88233,7 @@ } }, { - "accuracy": 0.9909241489673916, + "accuracy": 0.9909419746775376, "total_bits": 610188288, "q_proj": { "group_size": { @@ -88285,7 +88285,7 @@ } }, { - "accuracy": 0.9915860538419924, + "accuracy": 0.9916150554230339, "total_bits": 615184384, "q_proj": { "group_size": { @@ -88337,7 +88337,7 @@ } }, { - "accuracy": 0.9921938439733103, + "accuracy": 0.9921443438843677, "total_bits": 624115712, "q_proj": { "group_size": { @@ -88389,7 +88389,7 @@ } }, { - "accuracy": 0.9922348856925964, + "accuracy": 0.9923513766966368, "total_bits": 626637824, "q_proj": { "group_size": { @@ -88453,7 +88453,7 @@ } }, { - "accuracy": 0.9929556093717876, + "accuracy": 0.9930152516616019, "total_bits": 630519808, "q_proj": { "group_size": { @@ -88517,7 +88517,7 @@ } }, { - "accuracy": 0.9935751282855084, + "accuracy": 0.9935919728718305, "total_bits": 637526016, "q_proj": { "group_size": { @@ -88578,7 +88578,7 @@ } }, { - "accuracy": 0.9942007766742456, + "accuracy": 0.9941615254471177, "total_bits": 646987776, "q_proj": { "group_size": { @@ -88639,7 +88639,7 @@ } }, { - "accuracy": 0.9964970863963428, + "accuracy": 0.9965275898575783, "total_bits": 784904192, "q_proj": { "group_size": { @@ -88700,7 +88700,7 @@ } }, { - "accuracy": 0.9970847487841782, + "accuracy": 0.9970929442267669, "total_bits": 797982720, "q_proj": { "group_size": { @@ -88761,7 +88761,7 @@ } }, { - "accuracy": 0.9975781887769699, + "accuracy": 0.9976038621052316, "total_bits": 911912960, "q_proj": { "group_size": { @@ -88813,7 +88813,7 @@ } }, { - "accuracy": 0.9983488374242657, + "accuracy": 0.9983451899729276, "total_bits": 942882816, "q_proj": { "group_size": { @@ -88865,7 +88865,7 @@ } }, { - "accuracy": 0.9993156831417429, + "accuracy": 0.9993115611570446, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -88919,7 +88919,7 @@ ], "model.layers.45.mlp": [ { - "accuracy": 0.945914770427503, + "accuracy": 0.9460067937248632, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -88971,7 +88971,7 @@ } }, { - "accuracy": 0.9475420211490831, + "accuracy": 0.9476259131180613, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -89023,7 +89023,7 @@ } }, { - "accuracy": 0.9555790612572118, + "accuracy": 0.9556696509060106, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -89072,7 +89072,7 @@ } }, { - "accuracy": 0.9577777322969938, + "accuracy": 0.9578601216015062, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -89121,7 +89121,7 @@ } }, { - "accuracy": 0.9732583886698672, + "accuracy": 0.973315882055383, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -89173,7 +89173,7 @@ } }, { - "accuracy": 0.975185657802381, + "accuracy": 0.9752341838259446, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -89225,7 +89225,7 @@ } }, { - "accuracy": 0.9782630534548509, + "accuracy": 0.9783017509861996, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -89274,7 +89274,7 @@ } }, { - "accuracy": 0.9860967005553999, + "accuracy": 0.9861247257182473, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -89317,7 +89317,7 @@ } }, { - "accuracy": 0.9868392426716653, + "accuracy": 0.9868636884187397, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -89360,7 +89360,7 @@ } }, { - "accuracy": 0.9863438778801968, + "accuracy": 0.9863750040531158, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -89412,7 +89412,7 @@ } }, { - "accuracy": 0.988118547358011, + "accuracy": 0.9881416221982554, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -89464,7 +89464,7 @@ } }, { - "accuracy": 0.9930324327004584, + "accuracy": 0.9930439826689268, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -89516,7 +89516,7 @@ } }, { - "accuracy": 0.9940321210183596, + "accuracy": 0.9940430808224177, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -89568,7 +89568,7 @@ } }, { - "accuracy": 0.9961146351538206, + "accuracy": 0.9961184953388414, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -89611,7 +89611,7 @@ } }, { - "accuracy": 0.9964610695054656, + "accuracy": 0.9964637795561239, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -89660,7 +89660,7 @@ } }, { - "accuracy": 0.9970705497421717, + "accuracy": 0.997071663407903, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -89706,7 +89706,7 @@ } }, { - "accuracy": 0.9989007536910082, + "accuracy": 0.9988954538773549, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -89748,7 +89748,7 @@ ], "model.layers.46.self_attn": [ { - "accuracy": 0.9675373027199193, + "accuracy": 0.967542566751179, "total_bits": 320921600, "q_proj": { "group_size": { @@ -89812,7 +89812,7 @@ } }, { - "accuracy": 0.9688297196438438, + "accuracy": 0.9690181738451907, "total_bits": 329244672, "q_proj": { "group_size": { @@ -89876,7 +89876,7 @@ } }, { - "accuracy": 0.9711132582865263, + "accuracy": 0.971278965473175, "total_bits": 336188416, "q_proj": { "group_size": { @@ -89940,7 +89940,7 @@ } }, { - "accuracy": 0.9784119599743893, + "accuracy": 0.9786397814750671, "total_bits": 401721344, "q_proj": { "group_size": { @@ -90004,7 +90004,7 @@ } }, { - "accuracy": 0.9833266970358396, + "accuracy": 0.9833283769456964, "total_bits": 475443200, "q_proj": { "group_size": { @@ -90068,7 +90068,7 @@ } }, { - "accuracy": 0.9834891683176944, + "accuracy": 0.9834582633093784, "total_bits": 475642880, "q_proj": { "group_size": { @@ -90132,7 +90132,7 @@ } }, { - "accuracy": 0.9899188170307561, + "accuracy": 0.9899710291310361, "total_bits": 609923072, "q_proj": { "group_size": { @@ -90184,7 +90184,7 @@ } }, { - "accuracy": 0.990105905030903, + "accuracy": 0.9901424230713594, "total_bits": 610188288, "q_proj": { "group_size": { @@ -90236,7 +90236,7 @@ } }, { - "accuracy": 0.9910167902708054, + "accuracy": 0.9910369687958768, "total_bits": 615184384, "q_proj": { "group_size": { @@ -90288,7 +90288,7 @@ } }, { - "accuracy": 0.9915254916015425, + "accuracy": 0.9916296609138188, "total_bits": 624115712, "q_proj": { "group_size": { @@ -90340,7 +90340,7 @@ } }, { - "accuracy": 0.9915857911109924, + "accuracy": 0.9915825199139746, "total_bits": 626637824, "q_proj": { "group_size": { @@ -90404,7 +90404,7 @@ } }, { - "accuracy": 0.9922559480918082, + "accuracy": 0.992200673410767, "total_bits": 630519808, "q_proj": { "group_size": { @@ -90468,7 +90468,7 @@ } }, { - "accuracy": 0.9928960988396093, + "accuracy": 0.9928814800162065, "total_bits": 637526016, "q_proj": { "group_size": { @@ -90529,7 +90529,7 @@ } }, { - "accuracy": 0.9934790075609559, + "accuracy": 0.9934640624805501, "total_bits": 646987776, "q_proj": { "group_size": { @@ -90590,7 +90590,7 @@ } }, { - "accuracy": 0.996101213521079, + "accuracy": 0.9961132587570893, "total_bits": 784904192, "q_proj": { "group_size": { @@ -90651,7 +90651,7 @@ } }, { - "accuracy": 0.9966829995575704, + "accuracy": 0.9966907391422674, "total_bits": 797982720, "q_proj": { "group_size": { @@ -90712,7 +90712,7 @@ } }, { - "accuracy": 0.9973823400704485, + "accuracy": 0.9973838986142686, "total_bits": 911912960, "q_proj": { "group_size": { @@ -90764,7 +90764,7 @@ } }, { - "accuracy": 0.9981434186430354, + "accuracy": 0.9981222644840416, "total_bits": 942882816, "q_proj": { "group_size": { @@ -90816,7 +90816,7 @@ } }, { - "accuracy": 0.9992557435149425, + "accuracy": 0.9992603997846967, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -90870,7 +90870,7 @@ ], "model.layers.46.mlp": [ { - "accuracy": 0.9443400282608836, + "accuracy": 0.9444513069955927, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -90922,7 +90922,7 @@ } }, { - "accuracy": 0.9460537684591193, + "accuracy": 0.9461669921875, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -90974,7 +90974,7 @@ } }, { - "accuracy": 0.9544023149891904, + "accuracy": 0.9545114259970815, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -91023,7 +91023,7 @@ } }, { - "accuracy": 0.956669735281091, + "accuracy": 0.9567739398855912, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -91072,7 +91072,7 @@ } }, { - "accuracy": 0.9725739265743055, + "accuracy": 0.9726283644375048, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -91124,7 +91124,7 @@ } }, { - "accuracy": 0.9748171803198362, + "accuracy": 0.9748706613716326, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -91176,7 +91176,7 @@ } }, { - "accuracy": 0.9780096505817614, + "accuracy": 0.9780597435800653, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -91225,7 +91225,7 @@ } }, { - "accuracy": 0.9858066533741198, + "accuracy": 0.9858371395813791, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -91268,7 +91268,7 @@ } }, { - "accuracy": 0.9870266608501735, + "accuracy": 0.9870568882477911, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -91311,7 +91311,7 @@ } }, { - "accuracy": 0.9860685028527912, + "accuracy": 0.9860965970315432, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -91363,7 +91363,7 @@ } }, { - "accuracy": 0.9878184152276892, + "accuracy": 0.9878421629730024, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -91415,7 +91415,7 @@ } }, { - "accuracy": 0.9929037839174271, + "accuracy": 0.9929174208327344, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -91467,7 +91467,7 @@ } }, { - "accuracy": 0.993926758828916, + "accuracy": 0.9939408427790591, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -91519,7 +91519,7 @@ } }, { - "accuracy": 0.9961288669391682, + "accuracy": 0.9961358744062876, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -91562,7 +91562,7 @@ } }, { - "accuracy": 0.9963892494377337, + "accuracy": 0.9963945946411082, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -91611,7 +91611,7 @@ } }, { - "accuracy": 0.9970161252115902, + "accuracy": 0.9970190921112111, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -91657,7 +91657,7 @@ } }, { - "accuracy": 0.9989611362352183, + "accuracy": 0.998963000154809, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -91699,7 +91699,7 @@ ], "model.layers.47.self_attn": [ { - "accuracy": 0.962602016172911, + "accuracy": 0.9629354100478322, "total_bits": 320921600, "q_proj": { "group_size": { @@ -91763,7 +91763,7 @@ } }, { - "accuracy": 0.9676415794774106, + "accuracy": 0.9680395094971908, "total_bits": 329244672, "q_proj": { "group_size": { @@ -91827,7 +91827,7 @@ } }, { - "accuracy": 0.9703494216266432, + "accuracy": 0.9707411370779339, "total_bits": 336188416, "q_proj": { "group_size": { @@ -91891,7 +91891,7 @@ } }, { - "accuracy": 0.9774556363883772, + "accuracy": 0.9779215455055237, "total_bits": 401721344, "q_proj": { "group_size": { @@ -91955,7 +91955,7 @@ } }, { - "accuracy": 0.9828056128401506, + "accuracy": 0.9828746789380124, "total_bits": 475443200, "q_proj": { "group_size": { @@ -92019,7 +92019,7 @@ } }, { - "accuracy": 0.983063983289819, + "accuracy": 0.9831103698203438, "total_bits": 475642880, "q_proj": { "group_size": { @@ -92083,7 +92083,7 @@ } }, { - "accuracy": 0.9883185154513309, + "accuracy": 0.988415024782482, "total_bits": 609923072, "q_proj": { "group_size": { @@ -92135,7 +92135,7 @@ } }, { - "accuracy": 0.988618184861384, + "accuracy": 0.9886779808684399, "total_bits": 610188288, "q_proj": { "group_size": { @@ -92187,7 +92187,7 @@ } }, { - "accuracy": 0.9898987937914697, + "accuracy": 0.9899803290241643, "total_bits": 615184384, "q_proj": { "group_size": { @@ -92239,7 +92239,7 @@ } }, { - "accuracy": 0.9906725742314991, + "accuracy": 0.9908329201372046, "total_bits": 624115712, "q_proj": { "group_size": { @@ -92291,7 +92291,7 @@ } }, { - "accuracy": 0.9917076426117044, + "accuracy": 0.9917214018733878, "total_bits": 626637824, "q_proj": { "group_size": { @@ -92355,7 +92355,7 @@ } }, { - "accuracy": 0.9922810228247392, + "accuracy": 0.9922891996408764, "total_bits": 630519808, "q_proj": { "group_size": { @@ -92419,7 +92419,7 @@ } }, { - "accuracy": 0.9927174809731936, + "accuracy": 0.9927165445528532, "total_bits": 637526016, "q_proj": { "group_size": { @@ -92480,7 +92480,7 @@ } }, { - "accuracy": 0.9932820467572463, + "accuracy": 0.9933004191047267, "total_bits": 646987776, "q_proj": { "group_size": { @@ -92541,7 +92541,7 @@ } }, { - "accuracy": 0.9960053233723891, + "accuracy": 0.9960027427265519, "total_bits": 784904192, "q_proj": { "group_size": { @@ -92602,7 +92602,7 @@ } }, { - "accuracy": 0.9966263853405651, + "accuracy": 0.996623552551395, "total_bits": 797982720, "q_proj": { "group_size": { @@ -92663,7 +92663,7 @@ } }, { - "accuracy": 0.9972036306403185, + "accuracy": 0.997190506442597, "total_bits": 911912960, "q_proj": { "group_size": { @@ -92715,7 +92715,7 @@ } }, { - "accuracy": 0.9981076044863776, + "accuracy": 0.9981044877908731, "total_bits": 942882816, "q_proj": { "group_size": { @@ -92767,7 +92767,7 @@ } }, { - "accuracy": 0.9992036936608585, + "accuracy": 0.9992114203050733, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -92821,7 +92821,7 @@ ], "model.layers.47.mlp": [ { - "accuracy": 0.9432676654112966, + "accuracy": 0.9433886628401906, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -92873,7 +92873,7 @@ } }, { - "accuracy": 0.9450352380150243, + "accuracy": 0.9451398724003842, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -92925,7 +92925,7 @@ } }, { - "accuracy": 0.9537129308048048, + "accuracy": 0.9538255647609108, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -92974,7 +92974,7 @@ } }, { - "accuracy": 0.9560498940317255, + "accuracy": 0.9561616150956405, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -93023,7 +93023,7 @@ } }, { - "accuracy": 0.9720601219879953, + "accuracy": 0.9721210191124364, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -93075,7 +93075,7 @@ } }, { - "accuracy": 0.9743567595356389, + "accuracy": 0.9744079395344383, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -93127,7 +93127,7 @@ } }, { - "accuracy": 0.977694636897037, + "accuracy": 0.9777393623402244, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -93176,7 +93176,7 @@ } }, { - "accuracy": 0.985539088123723, + "accuracy": 0.985569494335275, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -93219,7 +93219,7 @@ } }, { - "accuracy": 0.9868034469453912, + "accuracy": 0.9868292306598864, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -93262,7 +93262,7 @@ } }, { - "accuracy": 0.9857935199612066, + "accuracy": 0.9858269597354689, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -93314,7 +93314,7 @@ } }, { - "accuracy": 0.9875904102074472, + "accuracy": 0.9876186243797603, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -93366,7 +93366,7 @@ } }, { - "accuracy": 0.9927585689645064, + "accuracy": 0.9927741987140555, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -93418,7 +93418,7 @@ } }, { - "accuracy": 0.9938086498724786, + "accuracy": 0.9938225620671323, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -93470,7 +93470,7 @@ } }, { - "accuracy": 0.9960583797410915, + "accuracy": 0.9960667839960048, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -93513,7 +93513,7 @@ } }, { - "accuracy": 0.9963057519573915, + "accuracy": 0.9963151432181659, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -93562,7 +93562,7 @@ } }, { - "accuracy": 0.9969685842332087, + "accuracy": 0.9969752426994475, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -93608,7 +93608,7 @@ } }, { - "accuracy": 0.9989426012494063, + "accuracy": 0.9989454012952352, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -93650,7 +93650,7 @@ ], "model.layers.48.self_attn": [ { - "accuracy": 0.9647048209842882, + "accuracy": 0.9649022911724291, "total_bits": 320921600, "q_proj": { "group_size": { @@ -93714,7 +93714,7 @@ } }, { - "accuracy": 0.9659538865089417, + "accuracy": 0.9660832756444028, "total_bits": 329244672, "q_proj": { "group_size": { @@ -93778,7 +93778,7 @@ } }, { - "accuracy": 0.9676503350860194, + "accuracy": 0.9679278888200459, "total_bits": 336188416, "q_proj": { "group_size": { @@ -93842,7 +93842,7 @@ } }, { - "accuracy": 0.9739121543733698, + "accuracy": 0.9741586980066801, "total_bits": 401721344, "q_proj": { "group_size": { @@ -93906,7 +93906,7 @@ } }, { - "accuracy": 0.9819551298492833, + "accuracy": 0.9818646829379233, "total_bits": 475443200, "q_proj": { "group_size": { @@ -93970,7 +93970,7 @@ } }, { - "accuracy": 0.9823086026467776, + "accuracy": 0.9820698656533894, "total_bits": 475642880, "q_proj": { "group_size": { @@ -94034,7 +94034,7 @@ } }, { - "accuracy": 0.9898032593099695, + "accuracy": 0.989793557085489, "total_bits": 609923072, "q_proj": { "group_size": { @@ -94086,7 +94086,7 @@ } }, { - "accuracy": 0.9902310151802866, + "accuracy": 0.9901768557335201, "total_bits": 610188288, "q_proj": { "group_size": { @@ -94138,7 +94138,7 @@ } }, { - "accuracy": 0.9909139699057529, + "accuracy": 0.9909208863973618, "total_bits": 615184384, "q_proj": { "group_size": { @@ -94190,7 +94190,7 @@ } }, { - "accuracy": 0.9912221282720566, + "accuracy": 0.9912604885665994, "total_bits": 624115712, "q_proj": { "group_size": { @@ -94242,7 +94242,7 @@ } }, { - "accuracy": 0.99104842073039, + "accuracy": 0.9909894074264326, "total_bits": 626637824, "q_proj": { "group_size": { @@ -94306,7 +94306,7 @@ } }, { - "accuracy": 0.9915867008660969, + "accuracy": 0.991700208500812, "total_bits": 630519808, "q_proj": { "group_size": { @@ -94370,7 +94370,7 @@ } }, { - "accuracy": 0.9925312666516555, + "accuracy": 0.992630695826129, "total_bits": 637526016, "q_proj": { "group_size": { @@ -94431,7 +94431,7 @@ } }, { - "accuracy": 0.9932273430259604, + "accuracy": 0.9932475560589841, "total_bits": 646987776, "q_proj": { "group_size": { @@ -94492,7 +94492,7 @@ } }, { - "accuracy": 0.9959951294095892, + "accuracy": 0.996022041690977, "total_bits": 784904192, "q_proj": { "group_size": { @@ -94553,7 +94553,7 @@ } }, { - "accuracy": 0.9966074506703176, + "accuracy": 0.9965947921338835, "total_bits": 797982720, "q_proj": { "group_size": { @@ -94614,7 +94614,7 @@ } }, { - "accuracy": 0.9973319547348901, + "accuracy": 0.9973557683590212, "total_bits": 911912960, "q_proj": { "group_size": { @@ -94666,7 +94666,7 @@ } }, { - "accuracy": 0.9982086906307622, + "accuracy": 0.9981947670641699, "total_bits": 942882816, "q_proj": { "group_size": { @@ -94718,7 +94718,7 @@ } }, { - "accuracy": 0.999255937279055, + "accuracy": 0.9992577670338122, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -94772,7 +94772,7 @@ ], "model.layers.48.mlp": [ { - "accuracy": 0.9419728517532349, + "accuracy": 0.942110582401878, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -94824,7 +94824,7 @@ } }, { - "accuracy": 0.9438734870207937, + "accuracy": 0.9439928845355385, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -94876,7 +94876,7 @@ } }, { - "accuracy": 0.9528889310987372, + "accuracy": 0.953000899992491, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -94925,7 +94925,7 @@ } }, { - "accuracy": 0.9553306698799133, + "accuracy": 0.9554373151377628, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -94974,7 +94974,7 @@ } }, { - "accuracy": 0.9714622309333399, + "accuracy": 0.9715311119430944, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -95026,7 +95026,7 @@ } }, { - "accuracy": 0.9738215568818545, + "accuracy": 0.9738725015991613, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -95078,7 +95078,7 @@ } }, { - "accuracy": 0.9773129852194535, + "accuracy": 0.9773583882733395, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -95127,7 +95127,7 @@ } }, { - "accuracy": 0.9851610346844322, + "accuracy": 0.985197492335972, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -95170,7 +95170,7 @@ } }, { - "accuracy": 0.9864678226019207, + "accuracy": 0.9864977347223383, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -95213,7 +95213,7 @@ } }, { - "accuracy": 0.9854861952756581, + "accuracy": 0.9855213290766666, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -95265,7 +95265,7 @@ } }, { - "accuracy": 0.9873284454408445, + "accuracy": 0.987357814845286, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -95317,7 +95317,7 @@ } }, { - "accuracy": 0.9925999555148577, + "accuracy": 0.9926189395942187, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -95369,7 +95369,7 @@ } }, { - "accuracy": 0.9936775263202818, + "accuracy": 0.9936915396859771, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -95421,7 +95421,7 @@ } }, { - "accuracy": 0.9959447705431989, + "accuracy": 0.9959536496745912, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -95464,7 +95464,7 @@ } }, { - "accuracy": 0.9962293925253969, + "accuracy": 0.99624134835444, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -95513,7 +95513,7 @@ } }, { - "accuracy": 0.9969152176850721, + "accuracy": 0.9969255214459017, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -95559,7 +95559,7 @@ } }, { - "accuracy": 0.998901172394031, + "accuracy": 0.9989043808493175, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -95601,7 +95601,7 @@ ], "model.layers.49.self_attn": [ { - "accuracy": 0.9654602533892581, + "accuracy": 0.9655440732052452, "total_bits": 320921600, "q_proj": { "group_size": { @@ -95665,7 +95665,7 @@ } }, { - "accuracy": 0.966844314023068, + "accuracy": 0.9669871894936812, "total_bits": 329244672, "q_proj": { "group_size": { @@ -95729,7 +95729,7 @@ } }, { - "accuracy": 0.969671434477756, + "accuracy": 0.9697374701499939, "total_bits": 336188416, "q_proj": { "group_size": { @@ -95793,7 +95793,7 @@ } }, { - "accuracy": 0.976446459167882, + "accuracy": 0.9764880145850935, "total_bits": 401721344, "q_proj": { "group_size": { @@ -95857,7 +95857,7 @@ } }, { - "accuracy": 0.9824358648375461, + "accuracy": 0.9824466626895102, "total_bits": 475443200, "q_proj": { "group_size": { @@ -95921,7 +95921,7 @@ } }, { - "accuracy": 0.9826659811170477, + "accuracy": 0.9826879548399072, "total_bits": 475642880, "q_proj": { "group_size": { @@ -95985,7 +95985,7 @@ } }, { - "accuracy": 0.9878488692798113, + "accuracy": 0.9878153400985819, "total_bits": 609923072, "q_proj": { "group_size": { @@ -96037,7 +96037,7 @@ } }, { - "accuracy": 0.9881475363907061, + "accuracy": 0.9881526898396643, "total_bits": 610188288, "q_proj": { "group_size": { @@ -96089,7 +96089,7 @@ } }, { - "accuracy": 0.9885023032364092, + "accuracy": 0.9890178496900358, "total_bits": 615184384, "q_proj": { "group_size": { @@ -96141,7 +96141,7 @@ } }, { - "accuracy": 0.9897380330060658, + "accuracy": 0.9897610819653461, "total_bits": 624115712, "q_proj": { "group_size": { @@ -96193,7 +96193,7 @@ } }, { - "accuracy": 0.9913009703159332, + "accuracy": 0.9912922241185841, "total_bits": 626637824, "q_proj": { "group_size": { @@ -96257,7 +96257,7 @@ } }, { - "accuracy": 0.9918790271407679, + "accuracy": 0.9918882360583857, "total_bits": 630519808, "q_proj": { "group_size": { @@ -96321,7 +96321,7 @@ } }, { - "accuracy": 0.9921572553484064, + "accuracy": 0.9921258515433261, "total_bits": 637526016, "q_proj": { "group_size": { @@ -96382,7 +96382,7 @@ } }, { - "accuracy": 0.9928326214614668, + "accuracy": 0.9928434883293352, "total_bits": 646987776, "q_proj": { "group_size": { @@ -96443,7 +96443,7 @@ } }, { - "accuracy": 0.995711941860224, + "accuracy": 0.9957353511923238, "total_bits": 784904192, "q_proj": { "group_size": { @@ -96504,7 +96504,7 @@ } }, { - "accuracy": 0.9962856726426828, + "accuracy": 0.9962856381347305, "total_bits": 797982720, "q_proj": { "group_size": { @@ -96565,7 +96565,7 @@ } }, { - "accuracy": 0.9968919506982753, + "accuracy": 0.9969165680047712, "total_bits": 911912960, "q_proj": { "group_size": { @@ -96617,7 +96617,7 @@ } }, { - "accuracy": 0.9979782159391203, + "accuracy": 0.997974610642383, "total_bits": 942882816, "q_proj": { "group_size": { @@ -96669,7 +96669,7 @@ } }, { - "accuracy": 0.9991373867776833, + "accuracy": 0.9991404551424479, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -96723,7 +96723,7 @@ ], "model.layers.49.mlp": [ { - "accuracy": 0.9416363615738719, + "accuracy": 0.9417492088518644, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -96775,7 +96775,7 @@ } }, { - "accuracy": 0.9435492942207738, + "accuracy": 0.9436646135229814, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -96827,7 +96827,7 @@ } }, { - "accuracy": 0.9528194477683619, + "accuracy": 0.9529204243107846, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -96876,7 +96876,7 @@ } }, { - "accuracy": 0.9553761137159247, + "accuracy": 0.9554706874646639, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -96925,7 +96925,7 @@ } }, { - "accuracy": 0.9712273603991458, + "accuracy": 0.971283473466572, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -96977,7 +96977,7 @@ } }, { - "accuracy": 0.9736142880038211, + "accuracy": 0.9736725499755458, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -97029,7 +97029,7 @@ } }, { - "accuracy": 0.9772356011365589, + "accuracy": 0.9772825303830599, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -97078,7 +97078,7 @@ } }, { - "accuracy": 0.9849111202516054, + "accuracy": 0.9849444047400826, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -97121,7 +97121,7 @@ } }, { - "accuracy": 0.9862448165291234, + "accuracy": 0.9862760760282215, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -97164,7 +97164,7 @@ } }, { - "accuracy": 0.9853858571303519, + "accuracy": 0.9854154178970739, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -97216,7 +97216,7 @@ } }, { - "accuracy": 0.9872046505150042, + "accuracy": 0.9872340622701143, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -97268,7 +97268,7 @@ } }, { - "accuracy": 0.992552513354703, + "accuracy": 0.9925669141505894, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -97320,7 +97320,7 @@ } }, { - "accuracy": 0.9936173628819617, + "accuracy": 0.9936306206019301, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -97372,7 +97372,7 @@ } }, { - "accuracy": 0.9958801351879772, + "accuracy": 0.9958899185845727, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -97415,7 +97415,7 @@ } }, { - "accuracy": 0.9962169555457014, + "accuracy": 0.9962259648661864, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -97464,7 +97464,7 @@ } }, { - "accuracy": 0.996951175951644, + "accuracy": 0.9969584520318007, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -97510,7 +97510,7 @@ } }, { - "accuracy": 0.9988861011439248, + "accuracy": 0.998889564193393, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -97552,7 +97552,7 @@ ], "model.layers.50.self_attn": [ { - "accuracy": 0.9613955083646273, + "accuracy": 0.9616266771366722, "total_bits": 320921600, "q_proj": { "group_size": { @@ -97616,7 +97616,7 @@ } }, { - "accuracy": 0.9627658191480135, + "accuracy": 0.9630704428020277, "total_bits": 329244672, "q_proj": { "group_size": { @@ -97680,7 +97680,7 @@ } }, { - "accuracy": 0.9655999698136982, + "accuracy": 0.9661006394185518, "total_bits": 336188416, "q_proj": { "group_size": { @@ -97744,7 +97744,7 @@ } }, { - "accuracy": 0.97285669100912, + "accuracy": 0.9734001285151431, "total_bits": 401721344, "q_proj": { "group_size": { @@ -97808,7 +97808,7 @@ } }, { - "accuracy": 0.9805299184824291, + "accuracy": 0.9805186186966143, "total_bits": 475443200, "q_proj": { "group_size": { @@ -97872,7 +97872,7 @@ } }, { - "accuracy": 0.980820706016139, + "accuracy": 0.9808006019968736, "total_bits": 475642880, "q_proj": { "group_size": { @@ -97936,7 +97936,7 @@ } }, { - "accuracy": 0.9866248573127546, + "accuracy": 0.986585342570355, "total_bits": 609923072, "q_proj": { "group_size": { @@ -97988,7 +97988,7 @@ } }, { - "accuracy": 0.9869832874912965, + "accuracy": 0.9869189246704704, "total_bits": 610188288, "q_proj": { "group_size": { @@ -98040,7 +98040,7 @@ } }, { - "accuracy": 0.9884902646667079, + "accuracy": 0.988621083529372, "total_bits": 615184384, "q_proj": { "group_size": { @@ -98092,7 +98092,7 @@ } }, { - "accuracy": 0.9889309437651383, + "accuracy": 0.9889737304888273, "total_bits": 624115712, "q_proj": { "group_size": { @@ -98144,7 +98144,7 @@ } }, { - "accuracy": 0.9904640795368898, + "accuracy": 0.9904082880208367, "total_bits": 626637824, "q_proj": { "group_size": { @@ -98208,7 +98208,7 @@ } }, { - "accuracy": 0.9910493461709273, + "accuracy": 0.9910518644671691, "total_bits": 630519808, "q_proj": { "group_size": { @@ -98272,7 +98272,7 @@ } }, { - "accuracy": 0.9914920322204891, + "accuracy": 0.9914831338744414, "total_bits": 637526016, "q_proj": { "group_size": { @@ -98333,7 +98333,7 @@ } }, { - "accuracy": 0.9922423833294919, + "accuracy": 0.992235320179086, "total_bits": 646987776, "q_proj": { "group_size": { @@ -98394,7 +98394,7 @@ } }, { - "accuracy": 0.9953052852498857, + "accuracy": 0.9952363642422777, "total_bits": 784904192, "q_proj": { "group_size": { @@ -98455,7 +98455,7 @@ } }, { - "accuracy": 0.9960311066947485, + "accuracy": 0.9960566292467871, "total_bits": 797982720, "q_proj": { "group_size": { @@ -98516,7 +98516,7 @@ } }, { - "accuracy": 0.9966094446809668, + "accuracy": 0.9965092614293098, "total_bits": 911912960, "q_proj": { "group_size": { @@ -98568,7 +98568,7 @@ } }, { - "accuracy": 0.997719656676054, + "accuracy": 0.9977593272924423, "total_bits": 942882816, "q_proj": { "group_size": { @@ -98620,7 +98620,7 @@ } }, { - "accuracy": 0.999075916644774, + "accuracy": 0.9990714158078557, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -98674,7 +98674,7 @@ ], "model.layers.50.mlp": [ { - "accuracy": 0.9410608567689595, + "accuracy": 0.9411726625342118, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -98726,7 +98726,7 @@ } }, { - "accuracy": 0.9429920284371627, + "accuracy": 0.9431037840090299, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -98778,7 +98778,7 @@ } }, { - "accuracy": 0.9525450436692489, + "accuracy": 0.9526417161289015, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -98827,7 +98827,7 @@ } }, { - "accuracy": 0.9552169849998072, + "accuracy": 0.9553062664835077, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -98876,7 +98876,7 @@ } }, { - "accuracy": 0.9708271779512104, + "accuracy": 0.9708844329181471, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -98928,7 +98928,7 @@ } }, { - "accuracy": 0.9732318583287691, + "accuracy": 0.9732831873391804, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -98980,7 +98980,7 @@ } }, { - "accuracy": 0.9770342848802868, + "accuracy": 0.9770784425108057, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -99029,7 +99029,7 @@ } }, { - "accuracy": 0.9846580969659906, + "accuracy": 0.9846885423911246, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -99072,7 +99072,7 @@ } }, { - "accuracy": 0.9860412949009946, + "accuracy": 0.9860700180656031, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -99115,7 +99115,7 @@ } }, { - "accuracy": 0.9851441414732682, + "accuracy": 0.9851696773579246, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -99167,7 +99167,7 @@ } }, { - "accuracy": 0.9870061992030394, + "accuracy": 0.9870286144708332, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -99219,7 +99219,7 @@ } }, { - "accuracy": 0.9924123600909585, + "accuracy": 0.9924301332549045, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -99271,7 +99271,7 @@ } }, { - "accuracy": 0.9935124806667629, + "accuracy": 0.9935264818762478, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -99323,7 +99323,7 @@ } }, { - "accuracy": 0.9958081802255229, + "accuracy": 0.9958190051348585, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -99366,7 +99366,7 @@ } }, { - "accuracy": 0.9961385746535502, + "accuracy": 0.9961482215868799, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -99415,7 +99415,7 @@ } }, { - "accuracy": 0.9969252008748682, + "accuracy": 0.9969324249970285, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -99461,7 +99461,7 @@ } }, { - "accuracy": 0.9988649904139733, + "accuracy": 0.9988690641169485, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -99503,7 +99503,7 @@ ], "model.layers.51.self_attn": [ { - "accuracy": 0.9581009965193898, + "accuracy": 0.957959953107332, "total_bits": 320921600, "q_proj": { "group_size": { @@ -99567,7 +99567,7 @@ } }, { - "accuracy": 0.9597636837708323, + "accuracy": 0.9594608890382867, "total_bits": 329244672, "q_proj": { "group_size": { @@ -99631,7 +99631,7 @@ } }, { - "accuracy": 0.9633042404526159, + "accuracy": 0.9628898532767045, "total_bits": 336188416, "q_proj": { "group_size": { @@ -99695,7 +99695,7 @@ } }, { - "accuracy": 0.9710493840669331, + "accuracy": 0.9704783088282535, "total_bits": 401721344, "q_proj": { "group_size": { @@ -99759,7 +99759,7 @@ } }, { - "accuracy": 0.9797418556715313, + "accuracy": 0.9800918666939986, "total_bits": 475443200, "q_proj": { "group_size": { @@ -99823,7 +99823,7 @@ } }, { - "accuracy": 0.9800206156153428, + "accuracy": 0.9803818997583891, "total_bits": 475642880, "q_proj": { "group_size": { @@ -99887,7 +99887,7 @@ } }, { - "accuracy": 0.9863794900869068, + "accuracy": 0.9868296463238565, "total_bits": 609923072, "q_proj": { "group_size": { @@ -99939,7 +99939,7 @@ } }, { - "accuracy": 0.9866785422751778, + "accuracy": 0.9871984382993296, "total_bits": 610188288, "q_proj": { "group_size": { @@ -99991,7 +99991,7 @@ } }, { - "accuracy": 0.9877052024791115, + "accuracy": 0.9882900550177223, "total_bits": 615184384, "q_proj": { "group_size": { @@ -100043,7 +100043,7 @@ } }, { - "accuracy": 0.988852897756978, + "accuracy": 0.988889817344515, "total_bits": 624115712, "q_proj": { "group_size": { @@ -100095,7 +100095,7 @@ } }, { - "accuracy": 0.9898128964399037, + "accuracy": 0.9897796095986116, "total_bits": 626637824, "q_proj": { "group_size": { @@ -100159,7 +100159,7 @@ } }, { - "accuracy": 0.9906295726173803, + "accuracy": 0.9906083729706312, "total_bits": 630519808, "q_proj": { "group_size": { @@ -100223,7 +100223,7 @@ } }, { - "accuracy": 0.990856431032482, + "accuracy": 0.9908230241976286, "total_bits": 637526016, "q_proj": { "group_size": { @@ -100284,7 +100284,7 @@ } }, { - "accuracy": 0.9917177738327729, + "accuracy": 0.9917977585604316, "total_bits": 646987776, "q_proj": { "group_size": { @@ -100345,7 +100345,7 @@ } }, { - "accuracy": 0.9950279130747444, + "accuracy": 0.9949694562115168, "total_bits": 784904192, "q_proj": { "group_size": { @@ -100406,7 +100406,7 @@ } }, { - "accuracy": 0.9958371492592912, + "accuracy": 0.9958578512856835, "total_bits": 797982720, "q_proj": { "group_size": { @@ -100467,7 +100467,7 @@ } }, { - "accuracy": 0.9964512390525717, + "accuracy": 0.9963703073169056, "total_bits": 911912960, "q_proj": { "group_size": { @@ -100519,7 +100519,7 @@ } }, { - "accuracy": 0.9976348192676118, + "accuracy": 0.9976293017205439, "total_bits": 942882816, "q_proj": { "group_size": { @@ -100571,7 +100571,7 @@ } }, { - "accuracy": 0.9989847615361214, + "accuracy": 0.9989968086347768, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -100625,7 +100625,7 @@ ], "model.layers.51.mlp": [ { - "accuracy": 0.9407902579558522, + "accuracy": 0.9408914540943346, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -100677,7 +100677,7 @@ } }, { - "accuracy": 0.9427834121804488, + "accuracy": 0.9428781145497372, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -100729,7 +100729,7 @@ } }, { - "accuracy": 0.9525916733239826, + "accuracy": 0.9526812559679935, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -100778,7 +100778,7 @@ } }, { - "accuracy": 0.9553482689355549, + "accuracy": 0.9554340996240315, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -100827,7 +100827,7 @@ } }, { - "accuracy": 0.9706874458413375, + "accuracy": 0.9707370086720115, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -100879,7 +100879,7 @@ } }, { - "accuracy": 0.973105932536878, + "accuracy": 0.9731521794670507, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -100931,7 +100931,7 @@ } }, { - "accuracy": 0.9770230148967943, + "accuracy": 0.9770604497507999, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -100980,7 +100980,7 @@ } }, { - "accuracy": 0.9845438646642786, + "accuracy": 0.9845718161055916, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -101023,7 +101023,7 @@ } }, { - "accuracy": 0.9859636127948761, + "accuracy": 0.9859886279231623, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -101066,7 +101066,7 @@ } }, { - "accuracy": 0.9850787347868869, + "accuracy": 0.9851056274614836, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -101118,7 +101118,7 @@ } }, { - "accuracy": 0.9869558748446012, + "accuracy": 0.9869799810020548, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -101170,7 +101170,7 @@ } }, { - "accuracy": 0.9923882202098244, + "accuracy": 0.9923997897850839, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -101222,7 +101222,7 @@ } }, { - "accuracy": 0.9934808584420305, + "accuracy": 0.9934925531086168, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -101274,7 +101274,7 @@ } }, { - "accuracy": 0.9958002551605827, + "accuracy": 0.9958080916028273, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -101317,7 +101317,7 @@ } }, { - "accuracy": 0.9961285265652757, + "accuracy": 0.996137445302386, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -101366,7 +101366,7 @@ } }, { - "accuracy": 0.9969260024004861, + "accuracy": 0.9969343997930226, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -101412,7 +101412,7 @@ } }, { - "accuracy": 0.9988715412389291, + "accuracy": 0.9988738986614504, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -101454,7 +101454,7 @@ ], "model.layers.52.self_attn": [ { - "accuracy": 0.960360376458419, + "accuracy": 0.9604806053011041, "total_bits": 320921600, "q_proj": { "group_size": { @@ -101518,7 +101518,7 @@ } }, { - "accuracy": 0.9621395217744928, + "accuracy": 0.9622661879188136, "total_bits": 329244672, "q_proj": { "group_size": { @@ -101582,7 +101582,7 @@ } }, { - "accuracy": 0.9651360637263248, + "accuracy": 0.9652300978961744, "total_bits": 336188416, "q_proj": { "group_size": { @@ -101646,7 +101646,7 @@ } }, { - "accuracy": 0.9738146835251859, + "accuracy": 0.973880968595806, "total_bits": 401721344, "q_proj": { "group_size": { @@ -101710,7 +101710,7 @@ } }, { - "accuracy": 0.9798088277641096, + "accuracy": 0.979855300564515, "total_bits": 475443200, "q_proj": { "group_size": { @@ -101774,7 +101774,7 @@ } }, { - "accuracy": 0.9802271911972448, + "accuracy": 0.9802797562197635, "total_bits": 475642880, "q_proj": { "group_size": { @@ -101838,7 +101838,7 @@ } }, { - "accuracy": 0.986567635285227, + "accuracy": 0.9866557560468975, "total_bits": 609923072, "q_proj": { "group_size": { @@ -101890,7 +101890,7 @@ } }, { - "accuracy": 0.9871156325465754, + "accuracy": 0.9871834587109717, "total_bits": 610188288, "q_proj": { "group_size": { @@ -101942,7 +101942,7 @@ } }, { - "accuracy": 0.9872933681073942, + "accuracy": 0.9873856259019751, "total_bits": 615184384, "q_proj": { "group_size": { @@ -101994,7 +101994,7 @@ } }, { - "accuracy": 0.9883708287226526, + "accuracy": 0.9884073961722223, "total_bits": 624115712, "q_proj": { "group_size": { @@ -102046,7 +102046,7 @@ } }, { - "accuracy": 0.9896534030374727, + "accuracy": 0.9896809552845202, "total_bits": 626637824, "q_proj": { "group_size": { @@ -102110,7 +102110,7 @@ } }, { - "accuracy": 0.9904652975107494, + "accuracy": 0.990468473026627, "total_bits": 630519808, "q_proj": { "group_size": { @@ -102174,7 +102174,7 @@ } }, { - "accuracy": 0.9907579288670891, + "accuracy": 0.990772407305868, "total_bits": 637526016, "q_proj": { "group_size": { @@ -102235,7 +102235,7 @@ } }, { - "accuracy": 0.9915162505287873, + "accuracy": 0.9914877297062623, "total_bits": 646987776, "q_proj": { "group_size": { @@ -102296,7 +102296,7 @@ } }, { - "accuracy": 0.9949922659679463, + "accuracy": 0.9950133169952192, "total_bits": 784904192, "q_proj": { "group_size": { @@ -102357,7 +102357,7 @@ } }, { - "accuracy": 0.9957376596959013, + "accuracy": 0.9957364844648462, "total_bits": 797982720, "q_proj": { "group_size": { @@ -102418,7 +102418,7 @@ } }, { - "accuracy": 0.9964129940459603, + "accuracy": 0.9964280779424467, "total_bits": 911912960, "q_proj": { "group_size": { @@ -102470,7 +102470,7 @@ } }, { - "accuracy": 0.9976944944968349, + "accuracy": 0.9976763637050202, "total_bits": 942882816, "q_proj": { "group_size": { @@ -102522,7 +102522,7 @@ } }, { - "accuracy": 0.9989685421123317, + "accuracy": 0.9989705790618533, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -102576,7 +102576,7 @@ ], "model.layers.52.mlp": [ { - "accuracy": 0.9383273626628675, + "accuracy": 0.9384277431588424, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -102628,7 +102628,7 @@ } }, { - "accuracy": 0.9404087317617316, + "accuracy": 0.9405117223137304, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -102680,7 +102680,7 @@ } }, { - "accuracy": 0.950473920295113, + "accuracy": 0.9505577212885806, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -102729,7 +102729,7 @@ } }, { - "accuracy": 0.9532826699708637, + "accuracy": 0.9533652255409643, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -102778,7 +102778,7 @@ } }, { - "accuracy": 0.9695011565559789, + "accuracy": 0.9695621917122289, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -102830,7 +102830,7 @@ } }, { - "accuracy": 0.9720169431284854, + "accuracy": 0.9720714500075892, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -102882,7 +102882,7 @@ } }, { - "accuracy": 0.9759826534672787, + "accuracy": 0.9760318367104781, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -102931,7 +102931,7 @@ } }, { - "accuracy": 0.983931690454483, + "accuracy": 0.983960531259838, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -102974,7 +102974,7 @@ } }, { - "accuracy": 0.9853865457208533, + "accuracy": 0.9854134352583634, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -103017,7 +103017,7 @@ } }, { - "accuracy": 0.9844766494474912, + "accuracy": 0.9845057945502432, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -103069,7 +103069,7 @@ } }, { - "accuracy": 0.9864270451821779, + "accuracy": 0.9864497561203805, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -103121,7 +103121,7 @@ } }, { - "accuracy": 0.992075840109273, + "accuracy": 0.9920913733934101, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -103173,7 +103173,7 @@ } }, { - "accuracy": 0.9932194007070441, + "accuracy": 0.9932319819927216, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -103225,7 +103225,7 @@ } }, { - "accuracy": 0.9956187727419954, + "accuracy": 0.9956266229089937, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -103268,7 +103268,7 @@ } }, { - "accuracy": 0.9959677167628941, + "accuracy": 0.9959747783447567, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -103317,7 +103317,7 @@ } }, { - "accuracy": 0.9967789898969626, + "accuracy": 0.9967855309185228, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -103363,7 +103363,7 @@ } }, { - "accuracy": 0.9988041213272434, + "accuracy": 0.9988046286529616, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -103405,7 +103405,7 @@ ], "model.layers.53.self_attn": [ { - "accuracy": 0.9590166618949488, + "accuracy": 0.9590088191785311, "total_bits": 320921600, "q_proj": { "group_size": { @@ -103469,7 +103469,7 @@ } }, { - "accuracy": 0.960692330410606, + "accuracy": 0.9606175987344039, "total_bits": 329244672, "q_proj": { "group_size": { @@ -103533,7 +103533,7 @@ } }, { - "accuracy": 0.9639503673503274, + "accuracy": 0.9639249506749605, "total_bits": 336188416, "q_proj": { "group_size": { @@ -103597,7 +103597,7 @@ } }, { - "accuracy": 0.9726763053944236, + "accuracy": 0.9725932384792128, "total_bits": 401721344, "q_proj": { "group_size": { @@ -103661,7 +103661,7 @@ } }, { - "accuracy": 0.9786694818421414, + "accuracy": 0.978767310318194, "total_bits": 475443200, "q_proj": { "group_size": { @@ -103725,7 +103725,7 @@ } }, { - "accuracy": 0.9789482151207171, + "accuracy": 0.9790611957248888, "total_bits": 475642880, "q_proj": { "group_size": { @@ -103789,7 +103789,7 @@ } }, { - "accuracy": 0.9857009225770047, + "accuracy": 0.9858001156857139, "total_bits": 609923072, "q_proj": { "group_size": { @@ -103841,7 +103841,7 @@ } }, { - "accuracy": 0.9860256345648515, + "accuracy": 0.9860998815611789, "total_bits": 610188288, "q_proj": { "group_size": { @@ -103893,7 +103893,7 @@ } }, { - "accuracy": 0.9870093527593111, + "accuracy": 0.987396198667978, "total_bits": 615184384, "q_proj": { "group_size": { @@ -103945,7 +103945,7 @@ } }, { - "accuracy": 0.9882750691551911, + "accuracy": 0.9884318003528997, "total_bits": 624115712, "q_proj": { "group_size": { @@ -103997,7 +103997,7 @@ } }, { - "accuracy": 0.9891238589035837, + "accuracy": 0.9891824879144367, "total_bits": 626637824, "q_proj": { "group_size": { @@ -104061,7 +104061,7 @@ } }, { - "accuracy": 0.9902044824863735, + "accuracy": 0.9902381081330148, "total_bits": 630519808, "q_proj": { "group_size": { @@ -104125,7 +104125,7 @@ } }, { - "accuracy": 0.9908215693737331, + "accuracy": 0.9908460927637, "total_bits": 637526016, "q_proj": { "group_size": { @@ -104186,7 +104186,7 @@ } }, { - "accuracy": 0.9914421643081465, + "accuracy": 0.9914823347016385, "total_bits": 646987776, "q_proj": { "group_size": { @@ -104247,7 +104247,7 @@ } }, { - "accuracy": 0.9949541989909975, + "accuracy": 0.9949656540625974, "total_bits": 784904192, "q_proj": { "group_size": { @@ -104308,7 +104308,7 @@ } }, { - "accuracy": 0.9956719098906768, + "accuracy": 0.9956942800628511, "total_bits": 797982720, "q_proj": { "group_size": { @@ -104369,7 +104369,7 @@ } }, { - "accuracy": 0.9964152994124513, + "accuracy": 0.9964293331692093, "total_bits": 911912960, "q_proj": { "group_size": { @@ -104421,7 +104421,7 @@ } }, { - "accuracy": 0.9975410876305479, + "accuracy": 0.9975577681080291, "total_bits": 942882816, "q_proj": { "group_size": { @@ -104473,7 +104473,7 @@ } }, { - "accuracy": 0.9990438331702822, + "accuracy": 0.9990395783005577, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -104527,7 +104527,7 @@ ], "model.layers.53.mlp": [ { - "accuracy": 0.9368705247577868, + "accuracy": 0.9369640601308722, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -104579,7 +104579,7 @@ } }, { - "accuracy": 0.9389301099275288, + "accuracy": 0.9390277109648052, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -104631,7 +104631,7 @@ } }, { - "accuracy": 0.9491041334051835, + "accuracy": 0.9492032622036181, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -104680,7 +104680,7 @@ } }, { - "accuracy": 0.9519333055144862, + "accuracy": 0.9520258621165627, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -104729,7 +104729,7 @@ } }, { - "accuracy": 0.9688395983294437, + "accuracy": 0.9688950620199505, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -104781,7 +104781,7 @@ } }, { - "accuracy": 0.9714017384930661, + "accuracy": 0.9714514581780684, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -104833,7 +104833,7 @@ } }, { - "accuracy": 0.975345923712379, + "accuracy": 0.9753855702124143, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -104882,7 +104882,7 @@ } }, { - "accuracy": 0.9836642004941639, + "accuracy": 0.9836904171266054, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -104925,7 +104925,7 @@ } }, { - "accuracy": 0.9851192737880506, + "accuracy": 0.9851454590496264, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -104968,7 +104968,7 @@ } }, { - "accuracy": 0.9841469541976327, + "accuracy": 0.9841728335932681, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -105020,7 +105020,7 @@ } }, { - "accuracy": 0.986138265383871, + "accuracy": 0.9861601842077155, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -105072,7 +105072,7 @@ } }, { - "accuracy": 0.9919170384344301, + "accuracy": 0.9919297930441404, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -105124,7 +105124,7 @@ } }, { - "accuracy": 0.9930810332298279, + "accuracy": 0.9930944395692725, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -105176,7 +105176,7 @@ } }, { - "accuracy": 0.9955550220451856, + "accuracy": 0.995561804034208, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -105219,7 +105219,7 @@ } }, { - "accuracy": 0.9958851651141518, + "accuracy": 0.9958912145934606, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -105268,7 +105268,7 @@ } }, { - "accuracy": 0.9967034184619, + "accuracy": 0.9967085542647462, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -105314,7 +105314,7 @@ } }, { - "accuracy": 0.9987959746075304, + "accuracy": 0.9987980723381042, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -105356,7 +105356,7 @@ ], "model.layers.54.self_attn": [ { - "accuracy": 0.9581252493356404, + "accuracy": 0.9579517245292664, "total_bits": 320921600, "q_proj": { "group_size": { @@ -105420,7 +105420,7 @@ } }, { - "accuracy": 0.9597993681305333, + "accuracy": 0.9596772758584273, "total_bits": 329244672, "q_proj": { "group_size": { @@ -105484,7 +105484,7 @@ } }, { - "accuracy": 0.9631577228244982, + "accuracy": 0.9629873852980764, "total_bits": 336188416, "q_proj": { "group_size": { @@ -105548,7 +105548,7 @@ } }, { - "accuracy": 0.9708959773967141, + "accuracy": 0.9706728019212422, "total_bits": 401721344, "q_proj": { "group_size": { @@ -105612,7 +105612,7 @@ } }, { - "accuracy": 0.9784045062567058, + "accuracy": 0.9783939264322582, "total_bits": 475443200, "q_proj": { "group_size": { @@ -105676,7 +105676,7 @@ } }, { - "accuracy": 0.9787917403798354, + "accuracy": 0.9788325177995783, "total_bits": 475642880, "q_proj": { "group_size": { @@ -105740,7 +105740,7 @@ } }, { - "accuracy": 0.9849602736924824, + "accuracy": 0.9849212561783037, "total_bits": 609923072, "q_proj": { "group_size": { @@ -105792,7 +105792,7 @@ } }, { - "accuracy": 0.9854598421799509, + "accuracy": 0.9854396597335213, "total_bits": 610188288, "q_proj": { "group_size": { @@ -105844,7 +105844,7 @@ } }, { - "accuracy": 0.9861629448438946, + "accuracy": 0.9860540173555675, "total_bits": 615184384, "q_proj": { "group_size": { @@ -105896,7 +105896,7 @@ } }, { - "accuracy": 0.9875596800917074, + "accuracy": 0.9873620695189426, "total_bits": 624115712, "q_proj": { "group_size": { @@ -105948,7 +105948,7 @@ } }, { - "accuracy": 0.9891763188337025, + "accuracy": 0.9891514182090759, "total_bits": 626637824, "q_proj": { "group_size": { @@ -106012,7 +106012,7 @@ } }, { - "accuracy": 0.9901662076774397, + "accuracy": 0.990211023311866, "total_bits": 630519808, "q_proj": { "group_size": { @@ -106076,7 +106076,7 @@ } }, { - "accuracy": 0.9905205933671248, + "accuracy": 0.9905148446559906, "total_bits": 637526016, "q_proj": { "group_size": { @@ -106137,7 +106137,7 @@ } }, { - "accuracy": 0.9913867068918127, + "accuracy": 0.9914352705604151, "total_bits": 646987776, "q_proj": { "group_size": { @@ -106198,7 +106198,7 @@ } }, { - "accuracy": 0.9948670440598538, + "accuracy": 0.9948381444341258, "total_bits": 784904192, "q_proj": { "group_size": { @@ -106259,7 +106259,7 @@ } }, { - "accuracy": 0.9956369470608862, + "accuracy": 0.9956142369069552, "total_bits": 797982720, "q_proj": { "group_size": { @@ -106320,7 +106320,7 @@ } }, { - "accuracy": 0.996267143440874, + "accuracy": 0.9962301120946282, "total_bits": 911912960, "q_proj": { "group_size": { @@ -106372,7 +106372,7 @@ } }, { - "accuracy": 0.997627326532414, + "accuracy": 0.9975716734403058, "total_bits": 942882816, "q_proj": { "group_size": { @@ -106424,7 +106424,7 @@ } }, { - "accuracy": 0.9989646198718172, + "accuracy": 0.9989706655278018, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -106478,7 +106478,7 @@ ], "model.layers.54.mlp": [ { - "accuracy": 0.9363208130786294, + "accuracy": 0.936419919917458, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -106530,7 +106530,7 @@ } }, { - "accuracy": 0.9382947808817813, + "accuracy": 0.9383886425118697, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -106582,7 +106582,7 @@ } }, { - "accuracy": 0.9484168761654904, + "accuracy": 0.948498547077179, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -106631,7 +106631,7 @@ } }, { - "accuracy": 0.9512926465586612, + "accuracy": 0.9513687742383856, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -106680,7 +106680,7 @@ } }, { - "accuracy": 0.9683664252883509, + "accuracy": 0.9684076434687564, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -106732,7 +106732,7 @@ } }, { - "accuracy": 0.9709638357162476, + "accuracy": 0.97099643004568, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -106784,7 +106784,7 @@ } }, { - "accuracy": 0.9749917670300132, + "accuracy": 0.9750215128848427, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -106833,7 +106833,7 @@ } }, { - "accuracy": 0.9834258807332892, + "accuracy": 0.9834513397593247, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -106876,7 +106876,7 @@ } }, { - "accuracy": 0.9849013560696652, + "accuracy": 0.9849218961439634, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -106919,7 +106919,7 @@ } }, { - "accuracy": 0.9839086218884117, + "accuracy": 0.9839310050010681, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -106971,7 +106971,7 @@ } }, { - "accuracy": 0.9859223554008886, + "accuracy": 0.9859406132447092, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -107023,7 +107023,7 @@ } }, { - "accuracy": 0.991793827790963, + "accuracy": 0.9918040327335659, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -107075,7 +107075,7 @@ } }, { - "accuracy": 0.9929735754665575, + "accuracy": 0.9929843646915335, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -107127,7 +107127,7 @@ } }, { - "accuracy": 0.9954889063772402, + "accuracy": 0.995495291132676, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -107170,7 +107170,7 @@ } }, { - "accuracy": 0.9958301268910107, + "accuracy": 0.9958332584876763, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -107219,7 +107219,7 @@ } }, { - "accuracy": 0.9966623226278707, + "accuracy": 0.9966651009101617, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -107265,7 +107265,7 @@ } }, { - "accuracy": 0.9987811006017422, + "accuracy": 0.9987818374249496, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -107307,7 +107307,7 @@ ], "model.layers.55.self_attn": [ { - "accuracy": 0.9601284045922129, + "accuracy": 0.9600867629051208, "total_bits": 320921600, "q_proj": { "group_size": { @@ -107371,7 +107371,7 @@ } }, { - "accuracy": 0.9616276715931139, + "accuracy": 0.9616104646732933, "total_bits": 329244672, "q_proj": { "group_size": { @@ -107435,7 +107435,7 @@ } }, { - "accuracy": 0.965016057616786, + "accuracy": 0.9649847400815863, "total_bits": 336188416, "q_proj": { "group_size": { @@ -107499,7 +107499,7 @@ } }, { - "accuracy": 0.9732846272619147, + "accuracy": 0.9732193946838379, "total_bits": 401721344, "q_proj": { "group_size": { @@ -107563,7 +107563,7 @@ } }, { - "accuracy": 0.9795080752749192, + "accuracy": 0.9795719040067572, "total_bits": 475443200, "q_proj": { "group_size": { @@ -107627,7 +107627,7 @@ } }, { - "accuracy": 0.980160349293759, + "accuracy": 0.9802679450888383, "total_bits": 475642880, "q_proj": { "group_size": { @@ -107691,7 +107691,7 @@ } }, { - "accuracy": 0.9868004902413017, + "accuracy": 0.9868481253322802, "total_bits": 609923072, "q_proj": { "group_size": { @@ -107743,7 +107743,7 @@ } }, { - "accuracy": 0.9877186715602875, + "accuracy": 0.9878556430339813, "total_bits": 610188288, "q_proj": { "group_size": { @@ -107795,7 +107795,7 @@ } }, { - "accuracy": 0.9886598210585745, + "accuracy": 0.988681790075804, "total_bits": 615184384, "q_proj": { "group_size": { @@ -107847,7 +107847,7 @@ } }, { - "accuracy": 0.989148736000061, + "accuracy": 0.9892819508125907, "total_bits": 624115712, "q_proj": { "group_size": { @@ -107899,7 +107899,7 @@ } }, { - "accuracy": 0.9899831335795554, + "accuracy": 0.9899889411110627, "total_bits": 626637824, "q_proj": { "group_size": { @@ -107963,7 +107963,7 @@ } }, { - "accuracy": 0.9905190381564593, + "accuracy": 0.9905316790467814, "total_bits": 630519808, "q_proj": { "group_size": { @@ -108027,7 +108027,7 @@ } }, { - "accuracy": 0.9912487362560473, + "accuracy": 0.9912655181006381, "total_bits": 637526016, "q_proj": { "group_size": { @@ -108088,7 +108088,7 @@ } }, { - "accuracy": 0.9917827703450856, + "accuracy": 0.9918025034038644, "total_bits": 646987776, "q_proj": { "group_size": { @@ -108149,7 +108149,7 @@ } }, { - "accuracy": 0.9949849342045031, + "accuracy": 0.9949717058947212, "total_bits": 784904192, "q_proj": { "group_size": { @@ -108210,7 +108210,7 @@ } }, { - "accuracy": 0.995901996367856, + "accuracy": 0.995904363299671, "total_bits": 797982720, "q_proj": { "group_size": { @@ -108271,7 +108271,7 @@ } }, { - "accuracy": 0.9964151962807304, + "accuracy": 0.9964025444106052, "total_bits": 911912960, "q_proj": { "group_size": { @@ -108323,7 +108323,7 @@ } }, { - "accuracy": 0.9976963555734408, + "accuracy": 0.9976916458261641, "total_bits": 942882816, "q_proj": { "group_size": { @@ -108375,7 +108375,7 @@ } }, { - "accuracy": 0.9990124034842378, + "accuracy": 0.9990071474917626, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -108429,7 +108429,7 @@ ], "model.layers.55.mlp": [ { - "accuracy": 0.9361953923576757, + "accuracy": 0.936297241010164, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -108481,7 +108481,7 @@ } }, { - "accuracy": 0.9381019441705001, + "accuracy": 0.938183866049114, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -108533,7 +108533,7 @@ } }, { - "accuracy": 0.9480595808280141, + "accuracy": 0.9481242487305089, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -108582,7 +108582,7 @@ } }, { - "accuracy": 0.9509225895530299, + "accuracy": 0.950988160936456, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -108631,7 +108631,7 @@ } }, { - "accuracy": 0.9682635508085552, + "accuracy": 0.9683135120492232, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -108683,7 +108683,7 @@ } }, { - "accuracy": 0.970856553629825, + "accuracy": 0.9708988415567499, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -108735,7 +108735,7 @@ } }, { - "accuracy": 0.9748459521092867, + "accuracy": 0.9748845445482355, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -108784,7 +108784,7 @@ } }, { - "accuracy": 0.9833760967380122, + "accuracy": 0.9834013832242865, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -108827,7 +108827,7 @@ } }, { - "accuracy": 0.9848611590109373, + "accuracy": 0.9848817273190147, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -108870,7 +108870,7 @@ } }, { - "accuracy": 0.9838621396767465, + "accuracy": 0.9838815739280299, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -108922,7 +108922,7 @@ } }, { - "accuracy": 0.9858845848786203, + "accuracy": 0.9859034026923933, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -108974,7 +108974,7 @@ } }, { - "accuracy": 0.9917710525424857, + "accuracy": 0.9917830856222856, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -109026,7 +109026,7 @@ } }, { - "accuracy": 0.9929530330394444, + "accuracy": 0.9929635038501338, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -109078,7 +109078,7 @@ } }, { - "accuracy": 0.9954775599272627, + "accuracy": 0.9954822757526448, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -109121,7 +109121,7 @@ } }, { - "accuracy": 0.9958194827562884, + "accuracy": 0.9958229645302421, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -109170,7 +109170,7 @@ } }, { - "accuracy": 0.9966433295293858, + "accuracy": 0.9966456756780022, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -109216,7 +109216,7 @@ } }, { - "accuracy": 0.9987697689548919, + "accuracy": 0.9987715638585781, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -109258,7 +109258,7 @@ ], "model.layers.56.self_attn": [ { - "accuracy": 0.9573582316699781, + "accuracy": 0.9573720317137868, "total_bits": 320921600, "q_proj": { "group_size": { @@ -109322,7 +109322,7 @@ } }, { - "accuracy": 0.9592474636278654, + "accuracy": 0.9592442826220864, "total_bits": 329244672, "q_proj": { "group_size": { @@ -109386,7 +109386,7 @@ } }, { - "accuracy": 0.9634109610005429, + "accuracy": 0.9633947171662983, "total_bits": 336188416, "q_proj": { "group_size": { @@ -109450,7 +109450,7 @@ } }, { - "accuracy": 0.9728977303755911, + "accuracy": 0.9728885293006897, "total_bits": 401721344, "q_proj": { "group_size": { @@ -109514,7 +109514,7 @@ } }, { - "accuracy": 0.9782423518205944, + "accuracy": 0.978207810928947, "total_bits": 475443200, "q_proj": { "group_size": { @@ -109578,7 +109578,7 @@ } }, { - "accuracy": 0.9785966151639035, + "accuracy": 0.978526350698973, "total_bits": 475642880, "q_proj": { "group_size": { @@ -109642,7 +109642,7 @@ } }, { - "accuracy": 0.9859853261395505, + "accuracy": 0.9859576758585478, "total_bits": 609923072, "q_proj": { "group_size": { @@ -109694,7 +109694,7 @@ } }, { - "accuracy": 0.9864034495855633, + "accuracy": 0.9863017970009854, "total_bits": 610188288, "q_proj": { "group_size": { @@ -109746,7 +109746,7 @@ } }, { - "accuracy": 0.987370648666432, + "accuracy": 0.9872357374743411, "total_bits": 615184384, "q_proj": { "group_size": { @@ -109798,7 +109798,7 @@ } }, { - "accuracy": 0.9881635127883208, + "accuracy": 0.9878378518317875, "total_bits": 624115712, "q_proj": { "group_size": { @@ -109850,7 +109850,7 @@ } }, { - "accuracy": 0.9890199460481343, + "accuracy": 0.9890048080369046, "total_bits": 626637824, "q_proj": { "group_size": { @@ -109914,7 +109914,7 @@ } }, { - "accuracy": 0.9898848463045923, + "accuracy": 0.9898632286410582, "total_bits": 630519808, "q_proj": { "group_size": { @@ -109978,7 +109978,7 @@ } }, { - "accuracy": 0.990474693085018, + "accuracy": 0.9904497752064153, "total_bits": 637526016, "q_proj": { "group_size": { @@ -110039,7 +110039,7 @@ } }, { - "accuracy": 0.9912459466018175, + "accuracy": 0.9911912852212003, "total_bits": 646987776, "q_proj": { "group_size": { @@ -110100,7 +110100,7 @@ } }, { - "accuracy": 0.9948746530633223, + "accuracy": 0.9948615384729285, "total_bits": 784904192, "q_proj": { "group_size": { @@ -110161,7 +110161,7 @@ } }, { - "accuracy": 0.9956453920979249, + "accuracy": 0.9956386469696697, "total_bits": 797982720, "q_proj": { "group_size": { @@ -110222,7 +110222,7 @@ } }, { - "accuracy": 0.9964208351938348, + "accuracy": 0.9964060489284364, "total_bits": 911912960, "q_proj": { "group_size": { @@ -110274,7 +110274,7 @@ } }, { - "accuracy": 0.9976381714406767, + "accuracy": 0.9976203296529619, "total_bits": 942882816, "q_proj": { "group_size": { @@ -110326,7 +110326,7 @@ } }, { - "accuracy": 0.9990166409039184, + "accuracy": 0.9990137327266367, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -110380,7 +110380,7 @@ ], "model.layers.56.mlp": [ { - "accuracy": 0.9352378406022724, + "accuracy": 0.9353189405642057, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -110432,7 +110432,7 @@ } }, { - "accuracy": 0.9371283869994314, + "accuracy": 0.9372134647871319, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -110484,7 +110484,7 @@ } }, { - "accuracy": 0.9473502510472348, + "accuracy": 0.9474093663065057, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -110533,7 +110533,7 @@ } }, { - "accuracy": 0.9503075041269001, + "accuracy": 0.9503640187414069, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -110582,7 +110582,7 @@ } }, { - "accuracy": 0.9679190892922251, + "accuracy": 0.967956687274732, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -110634,7 +110634,7 @@ } }, { - "accuracy": 0.9705210830035963, + "accuracy": 0.9705601303200972, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -110686,7 +110686,7 @@ } }, { - "accuracy": 0.9745830360211825, + "accuracy": 0.9746146186401969, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -110735,7 +110735,7 @@ } }, { - "accuracy": 0.9832244860498529, + "accuracy": 0.9832419455051422, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -110778,7 +110778,7 @@ } }, { - "accuracy": 0.9846991190784856, + "accuracy": 0.9847211837768555, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -110821,7 +110821,7 @@ } }, { - "accuracy": 0.9836927417077517, + "accuracy": 0.9837171576525036, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -110873,7 +110873,7 @@ } }, { - "accuracy": 0.9857385048740789, + "accuracy": 0.9857572819057264, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -110925,7 +110925,7 @@ } }, { - "accuracy": 0.9916942856813732, + "accuracy": 0.991707117933976, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -110977,7 +110977,7 @@ } }, { - "accuracy": 0.9928846280825766, + "accuracy": 0.9928941844325316, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -111029,7 +111029,7 @@ } }, { - "accuracy": 0.9954329878091812, + "accuracy": 0.9954387345596364, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -111072,7 +111072,7 @@ } }, { - "accuracy": 0.9957696003349203, + "accuracy": 0.995777077972889, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -111121,7 +111121,7 @@ } }, { - "accuracy": 0.9965972225917014, + "accuracy": 0.9966044473020654, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -111167,7 +111167,7 @@ } }, { - "accuracy": 0.9987325445797882, + "accuracy": 0.9987340138146752, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -111209,7 +111209,7 @@ ], "model.layers.57.self_attn": [ { - "accuracy": 0.9567002214883503, + "accuracy": 0.9566828263433356, "total_bits": 320921600, "q_proj": { "group_size": { @@ -111273,7 +111273,7 @@ } }, { - "accuracy": 0.9582353987191853, + "accuracy": 0.9582142704411557, "total_bits": 329244672, "q_proj": { "group_size": { @@ -111337,7 +111337,7 @@ } }, { - "accuracy": 0.9621693052743611, + "accuracy": 0.9621475370306718, "total_bits": 336188416, "q_proj": { "group_size": { @@ -111401,7 +111401,7 @@ } }, { - "accuracy": 0.9710918194369266, + "accuracy": 0.9710441608177988, "total_bits": 401721344, "q_proj": { "group_size": { @@ -111465,7 +111465,7 @@ } }, { - "accuracy": 0.97706065522997, + "accuracy": 0.9770960823485726, "total_bits": 475443200, "q_proj": { "group_size": { @@ -111529,7 +111529,7 @@ } }, { - "accuracy": 0.9773966199473331, + "accuracy": 0.9774525400839353, "total_bits": 475642880, "q_proj": { "group_size": { @@ -111593,7 +111593,7 @@ } }, { - "accuracy": 0.9841471502655431, + "accuracy": 0.9841347556365164, "total_bits": 609923072, "q_proj": { "group_size": { @@ -111645,7 +111645,7 @@ } }, { - "accuracy": 0.9845094821955028, + "accuracy": 0.9845217309500042, "total_bits": 610188288, "q_proj": { "group_size": { @@ -111697,7 +111697,7 @@ } }, { - "accuracy": 0.9871433855671632, + "accuracy": 0.9868911528273633, "total_bits": 615184384, "q_proj": { "group_size": { @@ -111749,7 +111749,7 @@ } }, { - "accuracy": 0.9876012888393904, + "accuracy": 0.9875201724077526, "total_bits": 624115712, "q_proj": { "group_size": { @@ -111801,7 +111801,7 @@ } }, { - "accuracy": 0.9886041691428736, + "accuracy": 0.9886628961876819, "total_bits": 626637824, "q_proj": { "group_size": { @@ -111865,7 +111865,7 @@ } }, { - "accuracy": 0.9894909152859136, + "accuracy": 0.9894968796717493, "total_bits": 630519808, "q_proj": { "group_size": { @@ -111929,7 +111929,7 @@ } }, { - "accuracy": 0.9899234387435412, + "accuracy": 0.9899873717835075, "total_bits": 637526016, "q_proj": { "group_size": { @@ -111990,7 +111990,7 @@ } }, { - "accuracy": 0.9909932770227131, + "accuracy": 0.9909822046756744, "total_bits": 646987776, "q_proj": { "group_size": { @@ -112051,7 +112051,7 @@ } }, { - "accuracy": 0.9945160950485029, + "accuracy": 0.9944745474740079, "total_bits": 784904192, "q_proj": { "group_size": { @@ -112112,7 +112112,7 @@ } }, { - "accuracy": 0.9954417242031348, + "accuracy": 0.9954385616277394, "total_bits": 797982720, "q_proj": { "group_size": { @@ -112173,7 +112173,7 @@ } }, { - "accuracy": 0.9959357776924184, + "accuracy": 0.9958759997236101, "total_bits": 911912960, "q_proj": { "group_size": { @@ -112225,7 +112225,7 @@ } }, { - "accuracy": 0.99752458420239, + "accuracy": 0.9975199279816527, "total_bits": 942882816, "q_proj": { "group_size": { @@ -112277,7 +112277,7 @@ } }, { - "accuracy": 0.9988947373471762, + "accuracy": 0.9988939065094057, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -112331,7 +112331,7 @@ ], "model.layers.57.mlp": [ { - "accuracy": 0.9339350712926764, + "accuracy": 0.9340059694490934, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -112383,7 +112383,7 @@ } }, { - "accuracy": 0.935798494439376, + "accuracy": 0.9358613114607961, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -112435,7 +112435,7 @@ } }, { - "accuracy": 0.9462451495622334, + "accuracy": 0.9462989631452059, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -112484,7 +112484,7 @@ } }, { - "accuracy": 0.9493032003703871, + "accuracy": 0.9493496982674849, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -112533,7 +112533,7 @@ } }, { - "accuracy": 0.9671492106036136, + "accuracy": 0.9671857545250341, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -112585,7 +112585,7 @@ } }, { - "accuracy": 0.9697916382237485, + "accuracy": 0.9698252834771809, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -112637,7 +112637,7 @@ } }, { - "accuracy": 0.9739995802703657, + "accuracy": 0.9740234644789445, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -112686,7 +112686,7 @@ } }, { - "accuracy": 0.9828700736949318, + "accuracy": 0.9828889464077196, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -112729,7 +112729,7 @@ } }, { - "accuracy": 0.9843945738516355, + "accuracy": 0.9844093667833429, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -112772,7 +112772,7 @@ } }, { - "accuracy": 0.9833100720455772, + "accuracy": 0.9833281024506217, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -112824,7 +112824,7 @@ } }, { - "accuracy": 0.98538065897791, + "accuracy": 0.9853973937662024, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -112876,7 +112876,7 @@ } }, { - "accuracy": 0.9914996224014383, + "accuracy": 0.9915080329305247, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -112928,7 +112928,7 @@ } }, { - "accuracy": 0.9927053177043011, + "accuracy": 0.9927131753218802, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -112980,7 +112980,7 @@ } }, { - "accuracy": 0.9953591298115881, + "accuracy": 0.9953632013578164, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -113023,7 +113023,7 @@ } }, { - "accuracy": 0.9956787785417155, + "accuracy": 0.9956819222945916, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -113072,7 +113072,7 @@ } }, { - "accuracy": 0.9965541088267377, + "accuracy": 0.9965563518436331, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -113118,7 +113118,7 @@ } }, { - "accuracy": 0.9987458907263843, + "accuracy": 0.9987474117231997, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -113160,7 +113160,7 @@ ], "model.layers.58.self_attn": [ { - "accuracy": 0.9570310398152, + "accuracy": 0.9571062640139931, "total_bits": 320921600, "q_proj": { "group_size": { @@ -113224,7 +113224,7 @@ } }, { - "accuracy": 0.9588187525146886, + "accuracy": 0.9588754992736013, "total_bits": 329244672, "q_proj": { "group_size": { @@ -113288,7 +113288,7 @@ } }, { - "accuracy": 0.9637406123311896, + "accuracy": 0.963786150279798, "total_bits": 336188416, "q_proj": { "group_size": { @@ -113352,7 +113352,7 @@ } }, { - "accuracy": 0.9726522972709254, + "accuracy": 0.9726871628510324, "total_bits": 401721344, "q_proj": { "group_size": { @@ -113416,7 +113416,7 @@ } }, { - "accuracy": 0.9782449210944929, + "accuracy": 0.9782546790022599, "total_bits": 475443200, "q_proj": { "group_size": { @@ -113480,7 +113480,7 @@ } }, { - "accuracy": 0.9787328949100093, + "accuracy": 0.9787489426763434, "total_bits": 475642880, "q_proj": { "group_size": { @@ -113544,7 +113544,7 @@ } }, { - "accuracy": 0.9860058019035741, + "accuracy": 0.9860169542463202, "total_bits": 609923072, "q_proj": { "group_size": { @@ -113596,7 +113596,7 @@ } }, { - "accuracy": 0.9866217092463845, + "accuracy": 0.9866073429584503, "total_bits": 610188288, "q_proj": { "group_size": { @@ -113648,7 +113648,7 @@ } }, { - "accuracy": 0.9878153040220863, + "accuracy": 0.9877691637528571, "total_bits": 615184384, "q_proj": { "group_size": { @@ -113700,7 +113700,7 @@ } }, { - "accuracy": 0.9886496427812075, + "accuracy": 0.9885973036289215, "total_bits": 624115712, "q_proj": { "group_size": { @@ -113752,7 +113752,7 @@ } }, { - "accuracy": 0.9893865389259238, + "accuracy": 0.989403613303837, "total_bits": 626637824, "q_proj": { "group_size": { @@ -113816,7 +113816,7 @@ } }, { - "accuracy": 0.9900242568630921, + "accuracy": 0.9900208092049548, "total_bits": 630519808, "q_proj": { "group_size": { @@ -113880,7 +113880,7 @@ } }, { - "accuracy": 0.9909472292975375, + "accuracy": 0.9909488841107017, "total_bits": 637526016, "q_proj": { "group_size": { @@ -113941,7 +113941,7 @@ } }, { - "accuracy": 0.9916425661036843, + "accuracy": 0.991643875837326, "total_bits": 646987776, "q_proj": { "group_size": { @@ -114002,7 +114002,7 @@ } }, { - "accuracy": 0.9949631859597406, + "accuracy": 0.9949788066901659, "total_bits": 784904192, "q_proj": { "group_size": { @@ -114063,7 +114063,7 @@ } }, { - "accuracy": 0.9958306103944778, + "accuracy": 0.9958335929795316, "total_bits": 797982720, "q_proj": { "group_size": { @@ -114124,7 +114124,7 @@ } }, { - "accuracy": 0.9964153519586513, + "accuracy": 0.9964370225605211, "total_bits": 911912960, "q_proj": { "group_size": { @@ -114176,7 +114176,7 @@ } }, { - "accuracy": 0.997708335322769, + "accuracy": 0.9977109749850473, "total_bits": 942882816, "q_proj": { "group_size": { @@ -114228,7 +114228,7 @@ } }, { - "accuracy": 0.9989972651789063, + "accuracy": 0.9989889465076359, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -114282,7 +114282,7 @@ ], "model.layers.58.mlp": [ { - "accuracy": 0.933761151213395, + "accuracy": 0.9338374075136686, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -114334,7 +114334,7 @@ } }, { - "accuracy": 0.935577668641743, + "accuracy": 0.9356592328924882, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -114386,7 +114386,7 @@ } }, { - "accuracy": 0.9458407038136533, + "accuracy": 0.945910604376542, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -114435,7 +114435,7 @@ } }, { - "accuracy": 0.9488802113031086, + "accuracy": 0.948944628238678, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -114484,7 +114484,7 @@ } }, { - "accuracy": 0.9670050332420751, + "accuracy": 0.9670524785393163, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -114536,7 +114536,7 @@ } }, { - "accuracy": 0.969669128719129, + "accuracy": 0.9697029935686212, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -114588,7 +114588,7 @@ } }, { - "accuracy": 0.9738307171746304, + "accuracy": 0.9738603108807614, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -114637,7 +114637,7 @@ } }, { - "accuracy": 0.9827891525469328, + "accuracy": 0.9828148264633981, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -114680,7 +114680,7 @@ } }, { - "accuracy": 0.9843283672081796, + "accuracy": 0.9843490860964123, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -114723,7 +114723,7 @@ } }, { - "accuracy": 0.983231782913208, + "accuracy": 0.9832515167562585, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -114775,7 +114775,7 @@ } }, { - "accuracy": 0.9853194669673317, + "accuracy": 0.9853359507886987, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -114827,7 +114827,7 @@ } }, { - "accuracy": 0.9914536766315761, + "accuracy": 0.9914665567247491, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -114879,7 +114879,7 @@ } }, { - "accuracy": 0.9926729108157911, + "accuracy": 0.9926806664780566, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -114931,7 +114931,7 @@ } }, { - "accuracy": 0.9953307438837854, + "accuracy": 0.9953388085490779, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -114974,7 +114974,7 @@ } }, { - "accuracy": 0.9956530003171218, + "accuracy": 0.9956568424638949, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -115023,7 +115023,7 @@ } }, { - "accuracy": 0.9965225603235396, + "accuracy": 0.9965255095770484, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -115069,7 +115069,7 @@ } }, { - "accuracy": 0.9987302273511887, + "accuracy": 0.9987304735144502, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -115111,7 +115111,7 @@ ], "model.layers.59.self_attn": [ { - "accuracy": 0.9571748093554848, + "accuracy": 0.9572336203173587, "total_bits": 320921600, "q_proj": { "group_size": { @@ -115175,7 +115175,7 @@ } }, { - "accuracy": 0.9587692492886594, + "accuracy": 0.9588479242826763, "total_bits": 329244672, "q_proj": { "group_size": { @@ -115239,7 +115239,7 @@ } }, { - "accuracy": 0.963310862842359, + "accuracy": 0.9633532574302272, "total_bits": 336188416, "q_proj": { "group_size": { @@ -115303,7 +115303,7 @@ } }, { - "accuracy": 0.9723703390673587, + "accuracy": 0.9723938672166121, "total_bits": 401721344, "q_proj": { "group_size": { @@ -115367,7 +115367,7 @@ } }, { - "accuracy": 0.9781908188995562, + "accuracy": 0.9781646195210909, "total_bits": 475443200, "q_proj": { "group_size": { @@ -115431,7 +115431,7 @@ } }, { - "accuracy": 0.9788288897589633, + "accuracy": 0.9787990827309457, "total_bits": 475642880, "q_proj": { "group_size": { @@ -115495,7 +115495,7 @@ } }, { - "accuracy": 0.985781519036544, + "accuracy": 0.9857676374284845, "total_bits": 609923072, "q_proj": { "group_size": { @@ -115547,7 +115547,7 @@ } }, { - "accuracy": 0.986615146461286, + "accuracy": 0.9866110855027249, "total_bits": 610188288, "q_proj": { "group_size": { @@ -115599,7 +115599,7 @@ } }, { - "accuracy": 0.9878571558939783, + "accuracy": 0.9878588648218858, "total_bits": 615184384, "q_proj": { "group_size": { @@ -115651,7 +115651,7 @@ } }, { - "accuracy": 0.9884801499153438, + "accuracy": 0.9884579518907949, "total_bits": 624115712, "q_proj": { "group_size": { @@ -115703,7 +115703,7 @@ } }, { - "accuracy": 0.9892241107790094, + "accuracy": 0.9892174915263527, "total_bits": 626637824, "q_proj": { "group_size": { @@ -115767,7 +115767,7 @@ } }, { - "accuracy": 0.9900756839074587, + "accuracy": 0.9900359480004561, "total_bits": 630519808, "q_proj": { "group_size": { @@ -115831,7 +115831,7 @@ } }, { - "accuracy": 0.9907327498260298, + "accuracy": 0.9906740557206305, "total_bits": 637526016, "q_proj": { "group_size": { @@ -115892,7 +115892,7 @@ } }, { - "accuracy": 0.9914746966801191, + "accuracy": 0.9915245692980917, "total_bits": 646987776, "q_proj": { "group_size": { @@ -115953,7 +115953,7 @@ } }, { - "accuracy": 0.9948690945380613, + "accuracy": 0.9948788230356417, "total_bits": 784904192, "q_proj": { "group_size": { @@ -116014,7 +116014,7 @@ } }, { - "accuracy": 0.9957549026922176, + "accuracy": 0.9957565849548892, "total_bits": 797982720, "q_proj": { "group_size": { @@ -116075,7 +116075,7 @@ } }, { - "accuracy": 0.9963475030503774, + "accuracy": 0.996354814422758, "total_bits": 911912960, "q_proj": { "group_size": { @@ -116127,7 +116127,7 @@ } }, { - "accuracy": 0.9976079756099927, + "accuracy": 0.9976134666878926, "total_bits": 942882816, "q_proj": { "group_size": { @@ -116179,7 +116179,7 @@ } }, { - "accuracy": 0.9989645944810227, + "accuracy": 0.998962095791572, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -116233,7 +116233,7 @@ ], "model.layers.59.mlp": [ { - "accuracy": 0.9357551524513646, + "accuracy": 0.935792464958994, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -116285,7 +116285,7 @@ } }, { - "accuracy": 0.9374575363962274, + "accuracy": 0.9375156415136237, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -116337,7 +116337,7 @@ } }, { - "accuracy": 0.9472153437765021, + "accuracy": 0.947236180305481, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -116386,7 +116386,7 @@ } }, { - "accuracy": 0.9501263718856008, + "accuracy": 0.9501430109927529, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -116435,7 +116435,7 @@ } }, { - "accuracy": 0.9679781229872453, + "accuracy": 0.9680195042961522, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -116487,7 +116487,7 @@ } }, { - "accuracy": 0.970545514633781, + "accuracy": 0.9705681142054106, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -116539,7 +116539,7 @@ } }, { - "accuracy": 0.9745393505221919, + "accuracy": 0.9745589808413857, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -116588,7 +116588,7 @@ } }, { - "accuracy": 0.9833203867862099, + "accuracy": 0.9833335123564068, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -116631,7 +116631,7 @@ } }, { - "accuracy": 0.9848044875421023, + "accuracy": 0.9848215219221617, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -116674,7 +116674,7 @@ } }, { - "accuracy": 0.9837433805591181, + "accuracy": 0.9837559743931419, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -116726,7 +116726,7 @@ } }, { - "accuracy": 0.9857610385668906, + "accuracy": 0.9857745358818456, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -116778,7 +116778,7 @@ } }, { - "accuracy": 0.9917222998644176, + "accuracy": 0.9917309543019847, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -116830,7 +116830,7 @@ } }, { - "accuracy": 0.9928974399441167, + "accuracy": 0.9929030529762569, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -116882,7 +116882,7 @@ } }, { - "accuracy": 0.9954798080419239, + "accuracy": 0.9954847403262791, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -116925,7 +116925,7 @@ } }, { - "accuracy": 0.9957918163977171, + "accuracy": 0.9957954373798872, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -116974,7 +116974,7 @@ } }, { - "accuracy": 0.9966083533669773, + "accuracy": 0.9966106438323071, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -117020,7 +117020,7 @@ } }, { - "accuracy": 0.9987558571523741, + "accuracy": 0.9987573556013798, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -117062,7 +117062,7 @@ ], "model.layers.60.self_attn": [ { - "accuracy": 0.9616448847871077, + "accuracy": 0.9616923394956087, "total_bits": 320921600, "q_proj": { "group_size": { @@ -117126,7 +117126,7 @@ } }, { - "accuracy": 0.9629194642368116, + "accuracy": 0.9630205537143507, "total_bits": 329244672, "q_proj": { "group_size": { @@ -117190,7 +117190,7 @@ } }, { - "accuracy": 0.96614114234322, + "accuracy": 0.9662236853649742, "total_bits": 336188416, "q_proj": { "group_size": { @@ -117254,7 +117254,7 @@ } }, { - "accuracy": 0.9737747437075565, + "accuracy": 0.9738738489778418, "total_bits": 401721344, "q_proj": { "group_size": { @@ -117318,7 +117318,7 @@ } }, { - "accuracy": 0.9801531732082367, + "accuracy": 0.980156725958774, "total_bits": 475443200, "q_proj": { "group_size": { @@ -117382,7 +117382,7 @@ } }, { - "accuracy": 0.9808755720916548, + "accuracy": 0.9808980352000186, "total_bits": 475642880, "q_proj": { "group_size": { @@ -117446,7 +117446,7 @@ } }, { - "accuracy": 0.987028023914287, + "accuracy": 0.9869414034642672, "total_bits": 609923072, "q_proj": { "group_size": { @@ -117498,7 +117498,7 @@ } }, { - "accuracy": 0.9879777094251231, + "accuracy": 0.9879973475870333, "total_bits": 610188288, "q_proj": { "group_size": { @@ -117550,7 +117550,7 @@ } }, { - "accuracy": 0.9892620012948388, + "accuracy": 0.9892579081811403, "total_bits": 615184384, "q_proj": { "group_size": { @@ -117602,7 +117602,7 @@ } }, { - "accuracy": 0.9892955406715995, + "accuracy": 0.9892734540136237, "total_bits": 624115712, "q_proj": { "group_size": { @@ -117654,7 +117654,7 @@ } }, { - "accuracy": 0.9902673010763369, + "accuracy": 0.9902800823512831, "total_bits": 626637824, "q_proj": { "group_size": { @@ -117718,7 +117718,7 @@ } }, { - "accuracy": 0.9909784558572268, + "accuracy": 0.9910046450401607, "total_bits": 630519808, "q_proj": { "group_size": { @@ -117782,7 +117782,7 @@ } }, { - "accuracy": 0.9913720237581354, + "accuracy": 0.9913821439994009, "total_bits": 637526016, "q_proj": { "group_size": { @@ -117843,7 +117843,7 @@ } }, { - "accuracy": 0.9921119801307979, + "accuracy": 0.9921318018122723, "total_bits": 646987776, "q_proj": { "group_size": { @@ -117904,7 +117904,7 @@ } }, { - "accuracy": 0.9950338206008861, + "accuracy": 0.9950397410675099, "total_bits": 784904192, "q_proj": { "group_size": { @@ -117965,7 +117965,7 @@ } }, { - "accuracy": 0.9960415265277812, + "accuracy": 0.9960421888451827, "total_bits": 797982720, "q_proj": { "group_size": { @@ -118026,7 +118026,7 @@ } }, { - "accuracy": 0.9963352499823821, + "accuracy": 0.9963346468774896, "total_bits": 911912960, "q_proj": { "group_size": { @@ -118078,7 +118078,7 @@ } }, { - "accuracy": 0.9977544471621513, + "accuracy": 0.9977544997083513, "total_bits": 942882816, "q_proj": { "group_size": { @@ -118130,7 +118130,7 @@ } }, { - "accuracy": 0.9989611233927702, + "accuracy": 0.9989744938517872, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -118184,7 +118184,7 @@ ], "model.layers.60.mlp": [ { - "accuracy": 0.935012698173523, + "accuracy": 0.9350637385719701, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -118236,7 +118236,7 @@ } }, { - "accuracy": 0.9367465533708271, + "accuracy": 0.9368264800623843, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -118288,7 +118288,7 @@ } }, { - "accuracy": 0.9463875105506495, + "accuracy": 0.9464483700300518, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -118337,7 +118337,7 @@ } }, { - "accuracy": 0.9492753242191515, + "accuracy": 0.9493322874370375, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -118386,7 +118386,7 @@ } }, { - "accuracy": 0.9677533507347107, + "accuracy": 0.9677803955580059, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -118438,7 +118438,7 @@ } }, { - "accuracy": 0.9703125294886137, + "accuracy": 0.9703511783951208, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -118490,7 +118490,7 @@ } }, { - "accuracy": 0.9741937533805245, + "accuracy": 0.9742300494721061, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -118539,7 +118539,7 @@ } }, { - "accuracy": 0.9832169460622888, + "accuracy": 0.9832374375117453, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -118582,7 +118582,7 @@ } }, { - "accuracy": 0.984705681863584, + "accuracy": 0.9847233248384375, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -118625,7 +118625,7 @@ } }, { - "accuracy": 0.9836111115781885, + "accuracy": 0.9836300909519196, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -118677,7 +118677,7 @@ } }, { - "accuracy": 0.98566094197725, + "accuracy": 0.9856770462111423, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -118729,7 +118729,7 @@ } }, { - "accuracy": 0.9916532510205319, + "accuracy": 0.9916624544482482, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -118781,7 +118781,7 @@ } }, { - "accuracy": 0.9928454427342666, + "accuracy": 0.9928565927241978, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -118833,7 +118833,7 @@ } }, { - "accuracy": 0.9954385859401602, + "accuracy": 0.9954448734459124, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -118876,7 +118876,7 @@ } }, { - "accuracy": 0.9957472274177953, + "accuracy": 0.9957516726694609, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -118925,7 +118925,7 @@ } }, { - "accuracy": 0.9965213321541485, + "accuracy": 0.9965247104042455, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -118971,7 +118971,7 @@ } }, { - "accuracy": 0.9987237425030846, + "accuracy": 0.998726995073651, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -119013,7 +119013,7 @@ ], "model.layers.61.self_attn": [ { - "accuracy": 0.9626173753487437, + "accuracy": 0.9626115184081228, "total_bits": 320921600, "q_proj": { "group_size": { @@ -119077,7 +119077,7 @@ } }, { - "accuracy": 0.963939651062614, + "accuracy": 0.963912662706877, "total_bits": 329244672, "q_proj": { "group_size": { @@ -119141,7 +119141,7 @@ } }, { - "accuracy": 0.9668720678279275, + "accuracy": 0.9668959441937899, "total_bits": 336188416, "q_proj": { "group_size": { @@ -119205,7 +119205,7 @@ } }, { - "accuracy": 0.9742786256890548, + "accuracy": 0.9742810647738608, "total_bits": 401721344, "q_proj": { "group_size": { @@ -119269,7 +119269,7 @@ } }, { - "accuracy": 0.9808291259564852, + "accuracy": 0.9809098824074394, "total_bits": 475443200, "q_proj": { "group_size": { @@ -119333,7 +119333,7 @@ } }, { - "accuracy": 0.9814355373382568, + "accuracy": 0.9814650981049788, "total_bits": 475642880, "q_proj": { "group_size": { @@ -119397,7 +119397,7 @@ } }, { - "accuracy": 0.987066061873185, + "accuracy": 0.9871683410908046, "total_bits": 609923072, "q_proj": { "group_size": { @@ -119449,7 +119449,7 @@ } }, { - "accuracy": 0.9878340316446204, + "accuracy": 0.9878990697233301, "total_bits": 610188288, "q_proj": { "group_size": { @@ -119501,7 +119501,7 @@ } }, { - "accuracy": 0.9889880308979436, + "accuracy": 0.98899565323403, "total_bits": 615184384, "q_proj": { "group_size": { @@ -119553,7 +119553,7 @@ } }, { - "accuracy": 0.9893045786179995, + "accuracy": 0.9892912193348533, "total_bits": 624115712, "q_proj": { "group_size": { @@ -119605,7 +119605,7 @@ } }, { - "accuracy": 0.9906117484757775, + "accuracy": 0.9906280840698042, "total_bits": 626637824, "q_proj": { "group_size": { @@ -119669,7 +119669,7 @@ } }, { - "accuracy": 0.9913441280001088, + "accuracy": 0.9913700881757235, "total_bits": 630519808, "q_proj": { "group_size": { @@ -119733,7 +119733,7 @@ } }, { - "accuracy": 0.9916982517430657, + "accuracy": 0.9917244362203699, "total_bits": 637526016, "q_proj": { "group_size": { @@ -119794,7 +119794,7 @@ } }, { - "accuracy": 0.9925273358821869, + "accuracy": 0.9925448502364912, "total_bits": 646987776, "q_proj": { "group_size": { @@ -119855,7 +119855,7 @@ } }, { - "accuracy": 0.9951142162868851, + "accuracy": 0.9951542313945921, "total_bits": 784904192, "q_proj": { "group_size": { @@ -119916,7 +119916,7 @@ } }, { - "accuracy": 0.9962613127733532, + "accuracy": 0.996276813902353, "total_bits": 797982720, "q_proj": { "group_size": { @@ -119977,7 +119977,7 @@ } }, { - "accuracy": 0.9963159525865003, + "accuracy": 0.996353321169552, "total_bits": 911912960, "q_proj": { "group_size": { @@ -120029,7 +120029,7 @@ } }, { - "accuracy": 0.9978776334931976, + "accuracy": 0.9978826345973894, "total_bits": 942882816, "q_proj": { "group_size": { @@ -120081,7 +120081,7 @@ } }, { - "accuracy": 0.999002513622767, + "accuracy": 0.9990054312113085, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -120135,7 +120135,7 @@ ], "model.layers.61.mlp": [ { - "accuracy": 0.9346254060142919, + "accuracy": 0.9346735226480585, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -120187,7 +120187,7 @@ } }, { - "accuracy": 0.9363490405835604, + "accuracy": 0.9364193301451833, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -120239,7 +120239,7 @@ } }, { - "accuracy": 0.9458207017497012, + "accuracy": 0.9458970960817839, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -120288,7 +120288,7 @@ } }, { - "accuracy": 0.9486844633754931, + "accuracy": 0.9487547843079818, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -120337,7 +120337,7 @@ } }, { - "accuracy": 0.9674971323264272, + "accuracy": 0.9675269879792866, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -120389,7 +120389,7 @@ } }, { - "accuracy": 0.9701052308082581, + "accuracy": 0.9701187861593146, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -120441,7 +120441,7 @@ } }, { - "accuracy": 0.973961917977584, + "accuracy": 0.9739748788507361, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -120490,7 +120490,7 @@ } }, { - "accuracy": 0.9830805078933114, + "accuracy": 0.9830981649850544, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -120533,7 +120533,7 @@ } }, { - "accuracy": 0.984599193460063, + "accuracy": 0.9846122719739613, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -120576,7 +120576,7 @@ } }, { - "accuracy": 0.9834979803938615, + "accuracy": 0.9835184938029239, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -120628,7 +120628,7 @@ } }, { - "accuracy": 0.9855574055721885, + "accuracy": 0.9855644106864929, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -120680,7 +120680,7 @@ } }, { - "accuracy": 0.9915942447750192, + "accuracy": 0.9916032568404549, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -120732,7 +120732,7 @@ } }, { - "accuracy": 0.9927892630037508, + "accuracy": 0.9927921585346523, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -120784,7 +120784,7 @@ } }, { - "accuracy": 0.9953954059042429, + "accuracy": 0.9954025863032592, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -120827,7 +120827,7 @@ } }, { - "accuracy": 0.995704929295339, + "accuracy": 0.9957091647543406, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -120876,7 +120876,7 @@ } }, { - "accuracy": 0.9964669233089999, + "accuracy": 0.9964706803622999, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -120922,7 +120922,7 @@ } }, { - "accuracy": 0.9986586118803212, + "accuracy": 0.9986604301160887, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -120964,7 +120964,7 @@ ], "model.layers.62.self_attn": [ { - "accuracy": 0.9623229691856786, + "accuracy": 0.9624576254894859, "total_bits": 320921600, "q_proj": { "group_size": { @@ -121028,7 +121028,7 @@ } }, { - "accuracy": 0.963591826589484, + "accuracy": 0.9637332652744494, "total_bits": 329244672, "q_proj": { "group_size": { @@ -121092,7 +121092,7 @@ } }, { - "accuracy": 0.9666535572001809, + "accuracy": 0.9667794610324659, "total_bits": 336188416, "q_proj": { "group_size": { @@ -121156,7 +121156,7 @@ } }, { - "accuracy": 0.9741578023684653, + "accuracy": 0.9743356328261527, "total_bits": 401721344, "q_proj": { "group_size": { @@ -121220,7 +121220,7 @@ } }, { - "accuracy": 0.9803918898105621, + "accuracy": 0.9804425114079526, "total_bits": 475443200, "q_proj": { "group_size": { @@ -121284,7 +121284,7 @@ } }, { - "accuracy": 0.9811616414471677, + "accuracy": 0.9811089399613833, "total_bits": 475642880, "q_proj": { "group_size": { @@ -121348,7 +121348,7 @@ } }, { - "accuracy": 0.9866360080869574, + "accuracy": 0.9866685679084376, "total_bits": 609923072, "q_proj": { "group_size": { @@ -121400,7 +121400,7 @@ } }, { - "accuracy": 0.9876505583524704, + "accuracy": 0.9875784281053042, "total_bits": 610188288, "q_proj": { "group_size": { @@ -121452,7 +121452,7 @@ } }, { - "accuracy": 0.9888162212936502, + "accuracy": 0.9888589884105482, "total_bits": 615184384, "q_proj": { "group_size": { @@ -121504,7 +121504,7 @@ } }, { - "accuracy": 0.9891508645132968, + "accuracy": 0.9891617925543534, "total_bits": 624115712, "q_proj": { "group_size": { @@ -121556,7 +121556,7 @@ } }, { - "accuracy": 0.9903567025535985, + "accuracy": 0.9903529325598165, "total_bits": 626637824, "q_proj": { "group_size": { @@ -121620,7 +121620,7 @@ } }, { - "accuracy": 0.9912442651234175, + "accuracy": 0.9912175410672238, "total_bits": 630519808, "q_proj": { "group_size": { @@ -121684,7 +121684,7 @@ } }, { - "accuracy": 0.9913775497361234, + "accuracy": 0.9913432637327596, "total_bits": 637526016, "q_proj": { "group_size": { @@ -121745,7 +121745,7 @@ } }, { - "accuracy": 0.9922854123931182, + "accuracy": 0.9922558955456081, "total_bits": 646987776, "q_proj": { "group_size": { @@ -121806,7 +121806,7 @@ } }, { - "accuracy": 0.9949409255855962, + "accuracy": 0.9949184264007368, "total_bits": 784904192, "q_proj": { "group_size": { @@ -121867,7 +121867,7 @@ } }, { - "accuracy": 0.9960532647214437, + "accuracy": 0.9960653962273347, "total_bits": 797982720, "q_proj": { "group_size": { @@ -121928,7 +121928,7 @@ } }, { - "accuracy": 0.9961393738263532, + "accuracy": 0.9961140967513386, "total_bits": 911912960, "q_proj": { "group_size": { @@ -121980,7 +121980,7 @@ } }, { - "accuracy": 0.9977638388150617, + "accuracy": 0.9977753495699481, "total_bits": 942882816, "q_proj": { "group_size": { @@ -122032,7 +122032,7 @@ } }, { - "accuracy": 0.9989789270452762, + "accuracy": 0.9989816238613505, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -122086,7 +122086,7 @@ ], "model.layers.62.mlp": [ { - "accuracy": 0.9352854615763614, + "accuracy": 0.935355964459871, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -122138,7 +122138,7 @@ } }, { - "accuracy": 0.9370312377026206, + "accuracy": 0.93709985833419, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -122190,7 +122190,7 @@ } }, { - "accuracy": 0.9460700122933638, + "accuracy": 0.9461258838051244, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -122239,7 +122239,7 @@ } }, { - "accuracy": 0.9488779526007802, + "accuracy": 0.9489293851350483, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -122288,7 +122288,7 @@ } }, { - "accuracy": 0.9678364521578738, + "accuracy": 0.9678723121944227, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -122340,7 +122340,7 @@ } }, { - "accuracy": 0.9703917879807321, + "accuracy": 0.9704261798607675, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -122392,7 +122392,7 @@ } }, { - "accuracy": 0.9740596303814336, + "accuracy": 0.9740834173403288, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -122441,7 +122441,7 @@ } }, { - "accuracy": 0.9832247997585096, + "accuracy": 0.9832455955053631, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -122484,7 +122484,7 @@ } }, { - "accuracy": 0.9847078997837869, + "accuracy": 0.9847306938547837, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -122527,7 +122527,7 @@ } }, { - "accuracy": 0.9836758924157996, + "accuracy": 0.9836906869160501, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -122579,7 +122579,7 @@ } }, { - "accuracy": 0.9856986623061331, + "accuracy": 0.9857173013059717, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -122631,7 +122631,7 @@ } }, { - "accuracy": 0.9916883040415613, + "accuracy": 0.9916967718224776, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -122683,7 +122683,7 @@ } }, { - "accuracy": 0.9928677356556842, + "accuracy": 0.9928748270398692, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -122735,7 +122735,7 @@ } }, { - "accuracy": 0.9954490179294034, + "accuracy": 0.9954557026687422, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -122778,7 +122778,7 @@ } }, { - "accuracy": 0.9957252109521314, + "accuracy": 0.9957302604850969, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -122827,7 +122827,7 @@ } }, { - "accuracy": 0.996445863654739, + "accuracy": 0.9964511241567763, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -122873,7 +122873,7 @@ } }, { - "accuracy": 0.9985896033283911, + "accuracy": 0.9985960203369981, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -122915,7 +122915,7 @@ ], "model.layers.63.self_attn": [ { - "accuracy": 0.9631868349878412, + "accuracy": 0.9631215208455136, "total_bits": 320921600, "q_proj": { "group_size": { @@ -122979,7 +122979,7 @@ } }, { - "accuracy": 0.9642837925961143, + "accuracy": 0.9642539620399475, "total_bits": 329244672, "q_proj": { "group_size": { @@ -123043,7 +123043,7 @@ } }, { - "accuracy": 0.9674978162112989, + "accuracy": 0.9674469138446607, "total_bits": 336188416, "q_proj": { "group_size": { @@ -123107,7 +123107,7 @@ } }, { - "accuracy": 0.9755316062977439, + "accuracy": 0.9754389618572435, "total_bits": 401721344, "q_proj": { "group_size": { @@ -123171,7 +123171,7 @@ } }, { - "accuracy": 0.9807167523785641, + "accuracy": 0.9807755053043365, "total_bits": 475443200, "q_proj": { "group_size": { @@ -123235,7 +123235,7 @@ } }, { - "accuracy": 0.9811421915104515, + "accuracy": 0.9812338085551011, "total_bits": 475642880, "q_proj": { "group_size": { @@ -123299,7 +123299,7 @@ } }, { - "accuracy": 0.9871099724581367, + "accuracy": 0.9871522737176794, "total_bits": 609923072, "q_proj": { "group_size": { @@ -123351,7 +123351,7 @@ } }, { - "accuracy": 0.9876212901190707, + "accuracy": 0.9877010811316339, "total_bits": 610188288, "q_proj": { "group_size": { @@ -123403,7 +123403,7 @@ } }, { - "accuracy": 0.9895829817182139, + "accuracy": 0.9896099834065688, "total_bits": 615184384, "q_proj": { "group_size": { @@ -123455,7 +123455,7 @@ } }, { - "accuracy": 0.9903275276485243, + "accuracy": 0.9903204033249303, "total_bits": 624115712, "q_proj": { "group_size": { @@ -123507,7 +123507,7 @@ } }, { - "accuracy": 0.9906569193852576, + "accuracy": 0.9906714746826574, "total_bits": 626637824, "q_proj": { "group_size": { @@ -123571,7 +123571,7 @@ } }, { - "accuracy": 0.9913816679465143, + "accuracy": 0.9913801550865173, "total_bits": 630519808, "q_proj": { "group_size": { @@ -123635,7 +123635,7 @@ } }, { - "accuracy": 0.9917507218687158, + "accuracy": 0.9917596798194083, "total_bits": 637526016, "q_proj": { "group_size": { @@ -123696,7 +123696,7 @@ } }, { - "accuracy": 0.9925645927065297, + "accuracy": 0.9925787268500579, "total_bits": 646987776, "q_proj": { "group_size": { @@ -123757,7 +123757,7 @@ } }, { - "accuracy": 0.9953135169650379, + "accuracy": 0.995320918920793, "total_bits": 784904192, "q_proj": { "group_size": { @@ -123818,7 +123818,7 @@ } }, { - "accuracy": 0.996311149706966, + "accuracy": 0.9963098411497316, "total_bits": 797982720, "q_proj": { "group_size": { @@ -123879,7 +123879,7 @@ } }, { - "accuracy": 0.9965522218691675, + "accuracy": 0.9965570851376182, "total_bits": 911912960, "q_proj": { "group_size": { @@ -123931,7 +123931,7 @@ } }, { - "accuracy": 0.9980311560395517, + "accuracy": 0.9980326253724726, "total_bits": 942882816, "q_proj": { "group_size": { @@ -123983,7 +123983,7 @@ } }, { - "accuracy": 0.9990556691037981, + "accuracy": 0.9990555821476799, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -124037,7 +124037,7 @@ ], "model.layers.63.mlp": [ { - "accuracy": 0.9348299001392565, + "accuracy": 0.9349082331908376, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -124089,7 +124089,7 @@ } }, { - "accuracy": 0.9365452026066027, + "accuracy": 0.9366291008497539, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -124141,7 +124141,7 @@ } }, { - "accuracy": 0.945646085237202, + "accuracy": 0.945699553740652, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -124190,7 +124190,7 @@ } }, { - "accuracy": 0.9484805496115434, + "accuracy": 0.9485284780201159, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -124239,7 +124239,7 @@ } }, { - "accuracy": 0.9675680367570174, + "accuracy": 0.9676138225354647, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -124291,7 +124291,7 @@ } }, { - "accuracy": 0.9701474403080187, + "accuracy": 0.9701854868939048, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -124343,7 +124343,7 @@ } }, { - "accuracy": 0.9738618072710539, + "accuracy": 0.9738959074020386, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -124392,7 +124392,7 @@ } }, { - "accuracy": 0.9830794083444696, + "accuracy": 0.9831011467858365, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -124435,7 +124435,7 @@ } }, { - "accuracy": 0.9846031665802002, + "accuracy": 0.9846210777759552, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -124478,7 +124478,7 @@ } }, { - "accuracy": 0.9835424046767386, + "accuracy": 0.9835647642612457, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -124530,7 +124530,7 @@ } }, { - "accuracy": 0.9855870871167434, + "accuracy": 0.9856070044793581, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -124582,7 +124582,7 @@ } }, { - "accuracy": 0.9916253529096904, + "accuracy": 0.9916363585936395, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -124634,7 +124634,7 @@ } }, { - "accuracy": 0.9928177148103714, + "accuracy": 0.992826980979819, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -124686,7 +124686,7 @@ } }, { - "accuracy": 0.9954221401559679, + "accuracy": 0.995428955476535, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -124729,7 +124729,7 @@ } }, { - "accuracy": 0.9957453388916818, + "accuracy": 0.9957488610556251, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -124778,7 +124778,7 @@ } }, { - "accuracy": 0.996489764828431, + "accuracy": 0.9964932756204354, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -124824,7 +124824,7 @@ } }, { - "accuracy": 0.998727126341117, + "accuracy": 0.9987290943727681, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -124866,7 +124866,7 @@ ], "model.layers.64.self_attn": [ { - "accuracy": 0.9673876291827151, + "accuracy": 0.9675218713910956, "total_bits": 320921600, "q_proj": { "group_size": { @@ -124930,7 +124930,7 @@ } }, { - "accuracy": 0.9683917070689955, + "accuracy": 0.9684902900143674, "total_bits": 329244672, "q_proj": { "group_size": { @@ -124994,7 +124994,7 @@ } }, { - "accuracy": 0.9715654034363596, + "accuracy": 0.9716033402242159, "total_bits": 336188416, "q_proj": { "group_size": { @@ -125058,7 +125058,7 @@ } }, { - "accuracy": 0.9785113240543165, + "accuracy": 0.978564695308083, "total_bits": 401721344, "q_proj": { "group_size": { @@ -125122,7 +125122,7 @@ } }, { - "accuracy": 0.9828911627593794, + "accuracy": 0.9829279090228834, "total_bits": 475443200, "q_proj": { "group_size": { @@ -125186,7 +125186,7 @@ } }, { - "accuracy": 0.9835346231335088, + "accuracy": 0.9836275624601465, "total_bits": 475642880, "q_proj": { "group_size": { @@ -125250,7 +125250,7 @@ } }, { - "accuracy": 0.9887270974485498, + "accuracy": 0.9887526396073794, "total_bits": 609923072, "q_proj": { "group_size": { @@ -125302,7 +125302,7 @@ } }, { - "accuracy": 0.9896010701593599, + "accuracy": 0.989707311517314, "total_bits": 610188288, "q_proj": { "group_size": { @@ -125354,7 +125354,7 @@ } }, { - "accuracy": 0.9908227810734197, + "accuracy": 0.9907993431154051, "total_bits": 615184384, "q_proj": { "group_size": { @@ -125406,7 +125406,7 @@ } }, { - "accuracy": 0.9910934896845567, + "accuracy": 0.99111434268324, "total_bits": 624115712, "q_proj": { "group_size": { @@ -125458,7 +125458,7 @@ } }, { - "accuracy": 0.9916660801360482, + "accuracy": 0.9916830870666002, "total_bits": 626637824, "q_proj": { "group_size": { @@ -125522,7 +125522,7 @@ } }, { - "accuracy": 0.992394145382078, + "accuracy": 0.9924065541279944, "total_bits": 630519808, "q_proj": { "group_size": { @@ -125586,7 +125586,7 @@ } }, { - "accuracy": 0.9927168535558801, + "accuracy": 0.9927289101638292, "total_bits": 637526016, "q_proj": { "group_size": { @@ -125647,7 +125647,7 @@ } }, { - "accuracy": 0.9934045554775941, + "accuracy": 0.9934207883320356, "total_bits": 646987776, "q_proj": { "group_size": { @@ -125708,7 +125708,7 @@ } }, { - "accuracy": 0.9958597453016984, + "accuracy": 0.9958677103644923, "total_bits": 784904192, "q_proj": { "group_size": { @@ -125769,7 +125769,7 @@ } }, { - "accuracy": 0.9966706013993213, + "accuracy": 0.9966811051494197, "total_bits": 797982720, "q_proj": { "group_size": { @@ -125830,7 +125830,7 @@ } }, { - "accuracy": 0.9970039335128508, + "accuracy": 0.9970116232962984, "total_bits": 911912960, "q_proj": { "group_size": { @@ -125882,7 +125882,7 @@ } }, { - "accuracy": 0.9981370731011817, + "accuracy": 0.9981515323253054, "total_bits": 942882816, "q_proj": { "group_size": { @@ -125934,7 +125934,7 @@ } }, { - "accuracy": 0.99916099109932, + "accuracy": 0.9991619450677383, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -125988,7 +125988,7 @@ ], "model.layers.64.mlp": [ { - "accuracy": 0.9371088931435033, + "accuracy": 0.9371873893235859, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -126040,7 +126040,7 @@ } }, { - "accuracy": 0.938796601797405, + "accuracy": 0.9388888384166517, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -126092,7 +126092,7 @@ } }, { - "accuracy": 0.9473026928148771, + "accuracy": 0.9473614002528944, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -126141,7 +126141,7 @@ } }, { - "accuracy": 0.9499564515916925, + "accuracy": 0.9500155699880499, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -126190,7 +126190,7 @@ } }, { - "accuracy": 0.9687132992242512, + "accuracy": 0.9687568796308417, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -126242,7 +126242,7 @@ } }, { - "accuracy": 0.9711821016512419, + "accuracy": 0.9712178926718863, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -126294,7 +126294,7 @@ } }, { - "accuracy": 0.9746838117900648, + "accuracy": 0.974718269548918, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -126343,7 +126343,7 @@ } }, { - "accuracy": 0.9836996119273337, + "accuracy": 0.9837192328352677, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -126386,7 +126386,7 @@ } }, { - "accuracy": 0.9851382578674116, + "accuracy": 0.9851492784525219, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -126429,7 +126429,7 @@ } }, { - "accuracy": 0.984128790466409, + "accuracy": 0.9841453934970655, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -126481,7 +126481,7 @@ } }, { - "accuracy": 0.9860894350629104, + "accuracy": 0.9861042154462714, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -126533,7 +126533,7 @@ } }, { - "accuracy": 0.9919225220617495, + "accuracy": 0.9919319419484389, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -126585,7 +126585,7 @@ } }, { - "accuracy": 0.993059618692649, + "accuracy": 0.9930681899974221, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -126637,7 +126637,7 @@ } }, { - "accuracy": 0.9955781843317183, + "accuracy": 0.9955859407782555, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -126680,7 +126680,7 @@ } }, { - "accuracy": 0.9958840710552115, + "accuracy": 0.9958905499232443, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -126729,7 +126729,7 @@ } }, { - "accuracy": 0.9965851593174433, + "accuracy": 0.9965900590545252, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -126775,7 +126775,7 @@ } }, { - "accuracy": 0.9987300154998114, + "accuracy": 0.9987315125763416, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -126817,7 +126817,7 @@ ], "model.layers.65.self_attn": [ { - "accuracy": 0.9700691511756495, + "accuracy": 0.9700785091048792, "total_bits": 320921600, "q_proj": { "group_size": { @@ -126881,7 +126881,7 @@ } }, { - "accuracy": 0.9710807957147297, + "accuracy": 0.9710868659772371, "total_bits": 329244672, "q_proj": { "group_size": { @@ -126945,7 +126945,7 @@ } }, { - "accuracy": 0.9734937109445271, + "accuracy": 0.9734962174766942, "total_bits": 336188416, "q_proj": { "group_size": { @@ -127009,7 +127009,7 @@ } }, { - "accuracy": 0.980129107048637, + "accuracy": 0.9801299242596877, "total_bits": 401721344, "q_proj": { "group_size": { @@ -127073,7 +127073,7 @@ } }, { - "accuracy": 0.9844275069864172, + "accuracy": 0.9844807496196345, "total_bits": 475443200, "q_proj": { "group_size": { @@ -127137,7 +127137,7 @@ } }, { - "accuracy": 0.9849112739688471, + "accuracy": 0.984960631320351, "total_bits": 475642880, "q_proj": { "group_size": { @@ -127201,7 +127201,7 @@ } }, { - "accuracy": 0.9897110971965288, + "accuracy": 0.9897623595438505, "total_bits": 609923072, "q_proj": { "group_size": { @@ -127253,7 +127253,7 @@ } }, { - "accuracy": 0.9903838195298847, + "accuracy": 0.9904458860033437, "total_bits": 610188288, "q_proj": { "group_size": { @@ -127305,7 +127305,7 @@ } }, { - "accuracy": 0.9911420713914069, + "accuracy": 0.9911453355299799, "total_bits": 615184384, "q_proj": { "group_size": { @@ -127357,7 +127357,7 @@ } }, { - "accuracy": 0.9919505331077074, + "accuracy": 0.9919735718714563, "total_bits": 624115712, "q_proj": { "group_size": { @@ -127409,7 +127409,7 @@ } }, { - "accuracy": 0.9924181817393554, + "accuracy": 0.992424853538212, "total_bits": 626637824, "q_proj": { "group_size": { @@ -127473,7 +127473,7 @@ } }, { - "accuracy": 0.9929383553956684, + "accuracy": 0.9929366401935878, "total_bits": 630519808, "q_proj": { "group_size": { @@ -127537,7 +127537,7 @@ } }, { - "accuracy": 0.9932805440927807, + "accuracy": 0.9932667471860585, "total_bits": 637526016, "q_proj": { "group_size": { @@ -127598,7 +127598,7 @@ } }, { - "accuracy": 0.9937437733537272, + "accuracy": 0.9937312457906572, "total_bits": 646987776, "q_proj": { "group_size": { @@ -127659,7 +127659,7 @@ } }, { - "accuracy": 0.9961401639800322, + "accuracy": 0.9961439551491487, "total_bits": 784904192, "q_proj": { "group_size": { @@ -127720,7 +127720,7 @@ } }, { - "accuracy": 0.9969090397812819, + "accuracy": 0.9969146204622168, "total_bits": 797982720, "q_proj": { "group_size": { @@ -127781,7 +127781,7 @@ } }, { - "accuracy": 0.9972215739912108, + "accuracy": 0.9972209595143795, "total_bits": 911912960, "q_proj": { "group_size": { @@ -127833,7 +127833,7 @@ } }, { - "accuracy": 0.9982836536670986, + "accuracy": 0.9982859019778276, "total_bits": 942882816, "q_proj": { "group_size": { @@ -127885,7 +127885,7 @@ } }, { - "accuracy": 0.9992202833098801, + "accuracy": 0.9992212972750789, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -127939,7 +127939,7 @@ ], "model.layers.65.mlp": [ { - "accuracy": 0.9376108332684165, + "accuracy": 0.9376628273411801, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -127991,7 +127991,7 @@ } }, { - "accuracy": 0.9392026788310001, + "accuracy": 0.9392831576497931, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -128043,7 +128043,7 @@ } }, { - "accuracy": 0.9477856598402324, + "accuracy": 0.9478514602309779, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -128092,7 +128092,7 @@ } }, { - "accuracy": 0.9504594708743849, + "accuracy": 0.9505179994984677, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -128141,7 +128141,7 @@ } }, { - "accuracy": 0.9689074095926786, + "accuracy": 0.968943388838517, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -128193,7 +128193,7 @@ } }, { - "accuracy": 0.9713293439463565, + "accuracy": 0.9713635601495442, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -128245,7 +128245,7 @@ } }, { - "accuracy": 0.9749010848371606, + "accuracy": 0.9749350296823602, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -128294,7 +128294,7 @@ } }, { - "accuracy": 0.983821114427165, + "accuracy": 0.9838462126882452, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -128337,7 +128337,7 @@ } }, { - "accuracy": 0.9852600740758997, + "accuracy": 0.9852779821345681, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -128380,7 +128380,7 @@ } }, { - "accuracy": 0.9842298940608376, + "accuracy": 0.9842499933744732, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -128432,7 +128432,7 @@ } }, { - "accuracy": 0.9861628193604318, + "accuracy": 0.986176671166169, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -128484,7 +128484,7 @@ } }, { - "accuracy": 0.9919819925960741, + "accuracy": 0.9919900470658353, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -128536,7 +128536,7 @@ } }, { - "accuracy": 0.9930983687701979, + "accuracy": 0.9931068075330634, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -128588,7 +128588,7 @@ } }, { - "accuracy": 0.9956286282915818, + "accuracy": 0.9956349871660534, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -128631,7 +128631,7 @@ } }, { - "accuracy": 0.9959264358407572, + "accuracy": 0.9959291631453916, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -128680,7 +128680,7 @@ } }, { - "accuracy": 0.9966334010425367, + "accuracy": 0.9966368953648367, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -128726,7 +128726,7 @@ } }, { - "accuracy": 0.9987814888162049, + "accuracy": 0.9987840397577536, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -128768,7 +128768,7 @@ ], "model.layers.66.self_attn": [ { - "accuracy": 0.9722122675494144, + "accuracy": 0.9722885018900821, "total_bits": 320921600, "q_proj": { "group_size": { @@ -128832,7 +128832,7 @@ } }, { - "accuracy": 0.973121304261057, + "accuracy": 0.9731819002251876, "total_bits": 329244672, "q_proj": { "group_size": { @@ -128896,7 +128896,7 @@ } }, { - "accuracy": 0.9753313488081882, + "accuracy": 0.9753953955675426, "total_bits": 336188416, "q_proj": { "group_size": { @@ -128960,7 +128960,7 @@ } }, { - "accuracy": 0.980905587735929, + "accuracy": 0.9809754694762983, "total_bits": 401721344, "q_proj": { "group_size": { @@ -129024,7 +129024,7 @@ } }, { - "accuracy": 0.9855589976436213, + "accuracy": 0.9856272198651966, "total_bits": 475443200, "q_proj": { "group_size": { @@ -129088,7 +129088,7 @@ } }, { - "accuracy": 0.9861516450580797, + "accuracy": 0.9861699813290646, "total_bits": 475642880, "q_proj": { "group_size": { @@ -129152,7 +129152,7 @@ } }, { - "accuracy": 0.9902368776108089, + "accuracy": 0.9902961787424589, "total_bits": 609923072, "q_proj": { "group_size": { @@ -129204,7 +129204,7 @@ } }, { - "accuracy": 0.991027693999441, + "accuracy": 0.9910388737916946, "total_bits": 610188288, "q_proj": { "group_size": { @@ -129256,7 +129256,7 @@ } }, { - "accuracy": 0.992249461381059, + "accuracy": 0.9922833027024018, "total_bits": 615184384, "q_proj": { "group_size": { @@ -129308,7 +129308,7 @@ } }, { - "accuracy": 0.9925689038477445, + "accuracy": 0.9926297021539587, "total_bits": 624115712, "q_proj": { "group_size": { @@ -129360,7 +129360,7 @@ } }, { - "accuracy": 0.9930271114173689, + "accuracy": 0.9930289301433062, "total_bits": 626637824, "q_proj": { "group_size": { @@ -129424,7 +129424,7 @@ } }, { - "accuracy": 0.9936099511228109, + "accuracy": 0.9936073512623185, "total_bits": 630519808, "q_proj": { "group_size": { @@ -129488,7 +129488,7 @@ } }, { - "accuracy": 0.9938383933744932, + "accuracy": 0.993815293829692, "total_bits": 637526016, "q_proj": { "group_size": { @@ -129549,7 +129549,7 @@ } }, { - "accuracy": 0.994376997806524, + "accuracy": 0.9943684366972823, "total_bits": 646987776, "q_proj": { "group_size": { @@ -129610,7 +129610,7 @@ } }, { - "accuracy": 0.9964501795015837, + "accuracy": 0.9964632152726776, "total_bits": 784904192, "q_proj": { "group_size": { @@ -129671,7 +129671,7 @@ } }, { - "accuracy": 0.9972152315864438, + "accuracy": 0.9972143502611863, "total_bits": 797982720, "q_proj": { "group_size": { @@ -129732,7 +129732,7 @@ } }, { - "accuracy": 0.9974074299006086, + "accuracy": 0.9974180532521323, "total_bits": 911912960, "q_proj": { "group_size": { @@ -129784,7 +129784,7 @@ } }, { - "accuracy": 0.9984526266588977, + "accuracy": 0.9984532040788939, "total_bits": 942882816, "q_proj": { "group_size": { @@ -129836,7 +129836,7 @@ } }, { - "accuracy": 0.9992678468360713, + "accuracy": 0.9992688762416181, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -129890,7 +129890,7 @@ ], "model.layers.66.mlp": [ { - "accuracy": 0.9386389757457533, + "accuracy": 0.9387047541768927, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -129942,7 +129942,7 @@ } }, { - "accuracy": 0.9402325718026412, + "accuracy": 0.9402926281878823, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -129994,7 +129994,7 @@ } }, { - "accuracy": 0.9485273549431249, + "accuracy": 0.9485762401631004, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -130043,7 +130043,7 @@ } }, { - "accuracy": 0.9510912393268786, + "accuracy": 0.9511407990204661, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -130092,7 +130092,7 @@ } }, { - "accuracy": 0.9693951042074906, + "accuracy": 0.9694353342056274, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -130144,7 +130144,7 @@ } }, { - "accuracy": 0.9717988967895508, + "accuracy": 0.9718246271735743, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -130196,7 +130196,7 @@ } }, { - "accuracy": 0.9752424767142848, + "accuracy": 0.9752655154780338, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -130245,7 +130245,7 @@ } }, { - "accuracy": 0.9840848947826185, + "accuracy": 0.9841088401643854, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -130288,7 +130288,7 @@ } }, { - "accuracy": 0.9854983875626012, + "accuracy": 0.9855126628750249, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -130331,7 +130331,7 @@ } }, { - "accuracy": 0.9844784391553778, + "accuracy": 0.9844955958818135, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -130383,7 +130383,7 @@ } }, { - "accuracy": 0.9863868465549067, + "accuracy": 0.9864007501225722, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -130435,7 +130435,7 @@ } }, { - "accuracy": 0.992105549887607, + "accuracy": 0.9921120405197144, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -130487,7 +130487,7 @@ } }, { - "accuracy": 0.9932061402421248, + "accuracy": 0.993213953940492, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -130539,7 +130539,7 @@ } }, { - "accuracy": 0.9956924389851721, + "accuracy": 0.9956996236976824, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -130582,7 +130582,7 @@ } }, { - "accuracy": 0.9959766864776611, + "accuracy": 0.9959827238007596, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -130631,7 +130631,7 @@ } }, { - "accuracy": 0.9966530721438559, + "accuracy": 0.9966588741854617, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -130677,7 +130677,7 @@ } }, { - "accuracy": 0.9987654820280639, + "accuracy": 0.9987686764644949, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -130719,7 +130719,7 @@ ], "model.layers.67.self_attn": [ { - "accuracy": 0.9764049759036616, + "accuracy": 0.9763436395871011, "total_bits": 320921600, "q_proj": { "group_size": { @@ -130783,7 +130783,7 @@ } }, { - "accuracy": 0.977181247974697, + "accuracy": 0.9771761251123328, "total_bits": 329244672, "q_proj": { "group_size": { @@ -130847,7 +130847,7 @@ } }, { - "accuracy": 0.9792466493029344, + "accuracy": 0.9792793879383489, "total_bits": 336188416, "q_proj": { "group_size": { @@ -130911,7 +130911,7 @@ } }, { - "accuracy": 0.9839906378796226, + "accuracy": 0.9840046904589, "total_bits": 401721344, "q_proj": { "group_size": { @@ -130975,7 +130975,7 @@ } }, { - "accuracy": 0.9876208195560857, + "accuracy": 0.9876286834478378, "total_bits": 475443200, "q_proj": { "group_size": { @@ -131039,7 +131039,7 @@ } }, { - "accuracy": 0.9881418708123659, + "accuracy": 0.9881806342225326, "total_bits": 475642880, "q_proj": { "group_size": { @@ -131103,7 +131103,7 @@ } }, { - "accuracy": 0.9915891297553715, + "accuracy": 0.9915706640795657, "total_bits": 609923072, "q_proj": { "group_size": { @@ -131155,7 +131155,7 @@ } }, { - "accuracy": 0.9922978438829121, + "accuracy": 0.9923096879532463, "total_bits": 610188288, "q_proj": { "group_size": { @@ -131207,7 +131207,7 @@ } }, { - "accuracy": 0.9932631183611719, + "accuracy": 0.9932688796206525, "total_bits": 615184384, "q_proj": { "group_size": { @@ -131259,7 +131259,7 @@ } }, { - "accuracy": 0.9937153631135037, + "accuracy": 0.9937190738947768, "total_bits": 624115712, "q_proj": { "group_size": { @@ -131311,7 +131311,7 @@ } }, { - "accuracy": 0.9939146535961252, + "accuracy": 0.9939292316374025, "total_bits": 626637824, "q_proj": { "group_size": { @@ -131375,7 +131375,7 @@ } }, { - "accuracy": 0.9944923304413494, + "accuracy": 0.9944993128117762, "total_bits": 630519808, "q_proj": { "group_size": { @@ -131439,7 +131439,7 @@ } }, { - "accuracy": 0.9945369911821265, + "accuracy": 0.9945497528502816, "total_bits": 637526016, "q_proj": { "group_size": { @@ -131500,7 +131500,7 @@ } }, { - "accuracy": 0.995180536649729, + "accuracy": 0.9951971616399916, "total_bits": 646987776, "q_proj": { "group_size": { @@ -131561,7 +131561,7 @@ } }, { - "accuracy": 0.9967892540520743, + "accuracy": 0.9967824597107736, "total_bits": 784904192, "q_proj": { "group_size": { @@ -131622,7 +131622,7 @@ } }, { - "accuracy": 0.9975839976809526, + "accuracy": 0.997582562855984, "total_bits": 797982720, "q_proj": { "group_size": { @@ -131683,7 +131683,7 @@ } }, { - "accuracy": 0.9975672648533395, + "accuracy": 0.9975523307527366, "total_bits": 911912960, "q_proj": { "group_size": { @@ -131735,7 +131735,7 @@ } }, { - "accuracy": 0.9986451525045069, + "accuracy": 0.998651232178274, "total_bits": 942882816, "q_proj": { "group_size": { @@ -131787,7 +131787,7 @@ } }, { - "accuracy": 0.9993341813648218, + "accuracy": 0.9993320959865263, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -131841,7 +131841,7 @@ ], "model.layers.67.mlp": [ { - "accuracy": 0.9395549422816226, + "accuracy": 0.9396588237662065, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -131893,7 +131893,7 @@ } }, { - "accuracy": 0.9411096196425588, + "accuracy": 0.9411814024573878, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -131945,7 +131945,7 @@ } }, { - "accuracy": 0.9491704953344244, + "accuracy": 0.9492418734650863, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -131994,7 +131994,7 @@ } }, { - "accuracy": 0.951630971933666, + "accuracy": 0.9517036425439935, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -132043,7 +132043,7 @@ } }, { - "accuracy": 0.9698232977013839, + "accuracy": 0.9698595655591864, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -132095,7 +132095,7 @@ } }, { - "accuracy": 0.9721885637233132, + "accuracy": 0.9722205588692113, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -132147,7 +132147,7 @@ } }, { - "accuracy": 0.975535943319923, + "accuracy": 0.9755650382292899, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -132196,7 +132196,7 @@ } }, { - "accuracy": 0.9843576487741972, + "accuracy": 0.9843749403953552, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -132239,7 +132239,7 @@ } }, { - "accuracy": 0.9857489796061265, + "accuracy": 0.9857689863757083, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -132282,7 +132282,7 @@ } }, { - "accuracy": 0.9847042670375422, + "accuracy": 0.984726745831339, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -132334,7 +132334,7 @@ } }, { - "accuracy": 0.9865871589434775, + "accuracy": 0.9865987849862952, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -132386,7 +132386,7 @@ } }, { - "accuracy": 0.9922313227465278, + "accuracy": 0.9922412767222053, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -132438,7 +132438,7 @@ } }, { - "accuracy": 0.9933210815254011, + "accuracy": 0.9933298810532218, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -132490,7 +132490,7 @@ } }, { - "accuracy": 0.9957877013244127, + "accuracy": 0.995793909618729, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -132533,7 +132533,7 @@ } }, { - "accuracy": 0.9960529647375408, + "accuracy": 0.9960576017436228, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -132582,7 +132582,7 @@ } }, { - "accuracy": 0.9967077923448462, + "accuracy": 0.9967124685645103, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -132628,7 +132628,7 @@ } }, { - "accuracy": 0.9988427888602018, + "accuracy": 0.9988446798371641, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -132670,7 +132670,7 @@ ], "model.layers.68.self_attn": [ { - "accuracy": 0.9742225628150137, + "accuracy": 0.9742597576818968, "total_bits": 320921600, "q_proj": { "group_size": { @@ -132734,7 +132734,7 @@ } }, { - "accuracy": 0.9749637732380315, + "accuracy": 0.9750134301813025, "total_bits": 329244672, "q_proj": { "group_size": { @@ -132798,7 +132798,7 @@ } }, { - "accuracy": 0.9772720070261705, + "accuracy": 0.9772458421556574, "total_bits": 336188416, "q_proj": { "group_size": { @@ -132862,7 +132862,7 @@ } }, { - "accuracy": 0.9822269238923725, + "accuracy": 0.9821424233286005, "total_bits": 401721344, "q_proj": { "group_size": { @@ -132926,7 +132926,7 @@ } }, { - "accuracy": 0.9863404474760357, + "accuracy": 0.9863587711986742, "total_bits": 475443200, "q_proj": { "group_size": { @@ -132990,7 +132990,7 @@ } }, { - "accuracy": 0.9868954608314916, + "accuracy": 0.9870035287580992, "total_bits": 475642880, "q_proj": { "group_size": { @@ -133054,7 +133054,7 @@ } }, { - "accuracy": 0.9903077459649036, + "accuracy": 0.9903352778208884, "total_bits": 609923072, "q_proj": { "group_size": { @@ -133106,7 +133106,7 @@ } }, { - "accuracy": 0.9910610765218735, + "accuracy": 0.9911864705775913, "total_bits": 610188288, "q_proj": { "group_size": { @@ -133158,7 +133158,7 @@ } }, { - "accuracy": 0.9920744919463208, + "accuracy": 0.9921044864152607, "total_bits": 615184384, "q_proj": { "group_size": { @@ -133210,7 +133210,7 @@ } }, { - "accuracy": 0.9924777785414144, + "accuracy": 0.9924494702565042, "total_bits": 624115712, "q_proj": { "group_size": { @@ -133262,7 +133262,7 @@ } }, { - "accuracy": 0.9932703783637599, + "accuracy": 0.9932607412338257, "total_bits": 626637824, "q_proj": { "group_size": { @@ -133326,7 +133326,7 @@ } }, { - "accuracy": 0.9938557120530229, + "accuracy": 0.9938642366936332, "total_bits": 630519808, "q_proj": { "group_size": { @@ -133390,7 +133390,7 @@ } }, { - "accuracy": 0.993936895932022, + "accuracy": 0.9939073088921999, "total_bits": 637526016, "q_proj": { "group_size": { @@ -133451,7 +133451,7 @@ } }, { - "accuracy": 0.994570084308323, + "accuracy": 0.9945847980285946, "total_bits": 646987776, "q_proj": { "group_size": { @@ -133512,7 +133512,7 @@ } }, { - "accuracy": 0.9963791429212219, + "accuracy": 0.99638855261238, "total_bits": 784904192, "q_proj": { "group_size": { @@ -133573,7 +133573,7 @@ } }, { - "accuracy": 0.9972283373537817, + "accuracy": 0.997227025267325, "total_bits": 797982720, "q_proj": { "group_size": { @@ -133634,7 +133634,7 @@ } }, { - "accuracy": 0.9971234743532381, + "accuracy": 0.9971299945916, "total_bits": 911912960, "q_proj": { "group_size": { @@ -133686,7 +133686,7 @@ } }, { - "accuracy": 0.9984933486894557, + "accuracy": 0.9984997791287146, "total_bits": 942882816, "q_proj": { "group_size": { @@ -133738,7 +133738,7 @@ } }, { - "accuracy": 0.9992231554106662, + "accuracy": 0.9992220825270602, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -133792,7 +133792,7 @@ ], "model.layers.68.mlp": [ { - "accuracy": 0.9400319111974615, + "accuracy": 0.9400904743294967, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -133844,7 +133844,7 @@ } }, { - "accuracy": 0.941580333207783, + "accuracy": 0.9416515513470298, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -133896,7 +133896,7 @@ } }, { - "accuracy": 0.9493884506978487, + "accuracy": 0.9494440398718181, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -133945,7 +133945,7 @@ } }, { - "accuracy": 0.9517879423342253, + "accuracy": 0.9518429291875738, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -133994,7 +133994,7 @@ } }, { - "accuracy": 0.9699929105608087, + "accuracy": 0.9700316378944799, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -134046,7 +134046,7 @@ } }, { - "accuracy": 0.9723423311584874, + "accuracy": 0.972375352131693, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -134098,7 +134098,7 @@ } }, { - "accuracy": 0.9755982035084775, + "accuracy": 0.97562785054508, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -134147,7 +134147,7 @@ } }, { - "accuracy": 0.9844467373270738, + "accuracy": 0.9844610408732766, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -134190,7 +134190,7 @@ } }, { - "accuracy": 0.9858201930397436, + "accuracy": 0.9858342471875643, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -134233,7 +134233,7 @@ } }, { - "accuracy": 0.9847787336299294, + "accuracy": 0.9847960032914814, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -134285,7 +134285,7 @@ } }, { - "accuracy": 0.9866492842373095, + "accuracy": 0.9866640348183481, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -134337,7 +134337,7 @@ } }, { - "accuracy": 0.9922619077720141, + "accuracy": 0.9922690603293871, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -134389,7 +134389,7 @@ } }, { - "accuracy": 0.9933453014022425, + "accuracy": 0.9933531896064156, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -134441,7 +134441,7 @@ } }, { - "accuracy": 0.9957971168976081, + "accuracy": 0.9958033326425051, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -134484,7 +134484,7 @@ } }, { - "accuracy": 0.9960681933321451, + "accuracy": 0.9960731228715495, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -134533,7 +134533,7 @@ } }, { - "accuracy": 0.9967074256978536, + "accuracy": 0.9967127118847872, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -134579,7 +134579,7 @@ } }, { - "accuracy": 0.9988530211542782, + "accuracy": 0.9988535040695417, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -134621,7 +134621,7 @@ ], "model.layers.69.self_attn": [ { - "accuracy": 0.9749151562389574, + "accuracy": 0.9749371440787065, "total_bits": 320921600, "q_proj": { "group_size": { @@ -134685,7 +134685,7 @@ } }, { - "accuracy": 0.9756422638893127, + "accuracy": 0.975688027708154, "total_bits": 329244672, "q_proj": { "group_size": { @@ -134749,7 +134749,7 @@ } }, { - "accuracy": 0.9780697979425129, + "accuracy": 0.9780842128552889, "total_bits": 336188416, "q_proj": { "group_size": { @@ -134813,7 +134813,7 @@ } }, { - "accuracy": 0.9837231416451303, + "accuracy": 0.983744026798951, "total_bits": 401721344, "q_proj": { "group_size": { @@ -134877,7 +134877,7 @@ } }, { - "accuracy": 0.9869938884910784, + "accuracy": 0.9869649080853713, "total_bits": 475443200, "q_proj": { "group_size": { @@ -134941,7 +134941,7 @@ } }, { - "accuracy": 0.9875515401363373, + "accuracy": 0.9874870777130127, "total_bits": 475642880, "q_proj": { "group_size": { @@ -135005,7 +135005,7 @@ } }, { - "accuracy": 0.9917615142307783, + "accuracy": 0.9916915062226748, "total_bits": 609923072, "q_proj": { "group_size": { @@ -135057,7 +135057,7 @@ } }, { - "accuracy": 0.9925444282983479, + "accuracy": 0.9924266612843463, "total_bits": 610188288, "q_proj": { "group_size": { @@ -135109,7 +135109,7 @@ } }, { - "accuracy": 0.993005342389408, + "accuracy": 0.9930235782736226, "total_bits": 615184384, "q_proj": { "group_size": { @@ -135161,7 +135161,7 @@ } }, { - "accuracy": 0.9933123737573624, + "accuracy": 0.993374767272096, "total_bits": 624115712, "q_proj": { "group_size": { @@ -135213,7 +135213,7 @@ } }, { - "accuracy": 0.9936598696206745, + "accuracy": 0.9936783023570713, "total_bits": 626637824, "q_proj": { "group_size": { @@ -135277,7 +135277,7 @@ } }, { - "accuracy": 0.9940884011356454, + "accuracy": 0.9940934992934528, "total_bits": 630519808, "q_proj": { "group_size": { @@ -135341,7 +135341,7 @@ } }, { - "accuracy": 0.994484372829136, + "accuracy": 0.9944896352918524, "total_bits": 637526016, "q_proj": { "group_size": { @@ -135402,7 +135402,7 @@ } }, { - "accuracy": 0.9949241021746084, + "accuracy": 0.9949282497951859, "total_bits": 646987776, "q_proj": { "group_size": { @@ -135463,7 +135463,7 @@ } }, { - "accuracy": 0.9969099373801759, + "accuracy": 0.9969154200271556, "total_bits": 784904192, "q_proj": { "group_size": { @@ -135524,7 +135524,7 @@ } }, { - "accuracy": 0.9974835383656778, + "accuracy": 0.9974829517304897, "total_bits": 797982720, "q_proj": { "group_size": { @@ -135585,7 +135585,7 @@ } }, { - "accuracy": 0.9978226776185789, + "accuracy": 0.9978261054738572, "total_bits": 911912960, "q_proj": { "group_size": { @@ -135637,7 +135637,7 @@ } }, { - "accuracy": 0.9986639547308809, + "accuracy": 0.9986640325698414, "total_bits": 942882816, "q_proj": { "group_size": { @@ -135689,7 +135689,7 @@ } }, { - "accuracy": 0.9993904731481483, + "accuracy": 0.9993945922898618, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -135743,7 +135743,7 @@ ], "model.layers.69.mlp": [ { - "accuracy": 0.9383871806295294, + "accuracy": 0.938458474058854, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -135795,7 +135795,7 @@ } }, { - "accuracy": 0.9399419332805433, + "accuracy": 0.9400156359923513, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -135847,7 +135847,7 @@ } }, { - "accuracy": 0.9479795129675614, + "accuracy": 0.9480186010661878, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -135896,7 +135896,7 @@ } }, { - "accuracy": 0.9504628212828385, + "accuracy": 0.9505008333607724, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -135945,7 +135945,7 @@ } }, { - "accuracy": 0.9691291984758879, + "accuracy": 0.969158062809392, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -135997,7 +135997,7 @@ } }, { - "accuracy": 0.9715472961726942, + "accuracy": 0.9715697828092074, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -136049,7 +136049,7 @@ } }, { - "accuracy": 0.974921570012444, + "accuracy": 0.9749437272548676, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -136098,7 +136098,7 @@ } }, { - "accuracy": 0.9839938471191808, + "accuracy": 0.9840040489246971, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -136141,7 +136141,7 @@ } }, { - "accuracy": 0.9854310249027453, + "accuracy": 0.9854408408466139, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -136184,7 +136184,7 @@ } }, { - "accuracy": 0.9843474375574213, + "accuracy": 0.9843623700894808, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -136236,7 +136236,7 @@ } }, { - "accuracy": 0.9862748557015469, + "accuracy": 0.9862885694754752, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -136288,7 +136288,7 @@ } }, { - "accuracy": 0.9920497936637778, + "accuracy": 0.9920568748524314, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -136340,7 +136340,7 @@ } }, { - "accuracy": 0.993164842066012, + "accuracy": 0.9931721138326746, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -136392,7 +136392,7 @@ } }, { - "accuracy": 0.995698980987072, + "accuracy": 0.9957039293489958, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -136435,7 +136435,7 @@ } }, { - "accuracy": 0.9959615504271105, + "accuracy": 0.9959627868313539, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -136484,7 +136484,7 @@ } }, { - "accuracy": 0.9966254900944861, + "accuracy": 0.9966264970992741, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -136530,7 +136530,7 @@ } }, { - "accuracy": 0.9988260349552882, + "accuracy": 0.9988259269218696, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -136572,7 +136572,7 @@ ], "model.layers.70.self_attn": [ { - "accuracy": 0.9729410943232084, + "accuracy": 0.9728989757989582, "total_bits": 320921600, "q_proj": { "group_size": { @@ -136636,7 +136636,7 @@ } }, { - "accuracy": 0.97388903404537, + "accuracy": 0.9738847048659074, "total_bits": 329244672, "q_proj": { "group_size": { @@ -136700,7 +136700,7 @@ } }, { - "accuracy": 0.9765033737609261, + "accuracy": 0.9764793734801444, "total_bits": 336188416, "q_proj": { "group_size": { @@ -136764,7 +136764,7 @@ } }, { - "accuracy": 0.9822052968175787, + "accuracy": 0.9821657710953763, "total_bits": 401721344, "q_proj": { "group_size": { @@ -136828,7 +136828,7 @@ } }, { - "accuracy": 0.9857922886547289, + "accuracy": 0.9857936328963229, "total_bits": 475443200, "q_proj": { "group_size": { @@ -136892,7 +136892,7 @@ } }, { - "accuracy": 0.9863715062015935, + "accuracy": 0.9863653010443637, "total_bits": 475642880, "q_proj": { "group_size": { @@ -136956,7 +136956,7 @@ } }, { - "accuracy": 0.9903415841491598, + "accuracy": 0.9903539717197418, "total_bits": 609923072, "q_proj": { "group_size": { @@ -137008,7 +137008,7 @@ } }, { - "accuracy": 0.9911219626665115, + "accuracy": 0.991114370917019, "total_bits": 610188288, "q_proj": { "group_size": { @@ -137060,7 +137060,7 @@ } }, { - "accuracy": 0.9920979165717175, + "accuracy": 0.9920989392619384, "total_bits": 615184384, "q_proj": { "group_size": { @@ -137112,7 +137112,7 @@ } }, { - "accuracy": 0.9926125454275232, + "accuracy": 0.9925986379384995, "total_bits": 624115712, "q_proj": { "group_size": { @@ -137164,7 +137164,7 @@ } }, { - "accuracy": 0.9930518724416432, + "accuracy": 0.9930574156736073, "total_bits": 626637824, "q_proj": { "group_size": { @@ -137228,7 +137228,7 @@ } }, { - "accuracy": 0.9936560325716671, + "accuracy": 0.9936652254117163, "total_bits": 630519808, "q_proj": { "group_size": { @@ -137292,7 +137292,7 @@ } }, { - "accuracy": 0.9938749439622226, + "accuracy": 0.9938674631871676, "total_bits": 637526016, "q_proj": { "group_size": { @@ -137353,7 +137353,7 @@ } }, { - "accuracy": 0.9945449005616339, + "accuracy": 0.9945597374125531, "total_bits": 646987776, "q_proj": { "group_size": { @@ -137414,7 +137414,7 @@ } }, { - "accuracy": 0.9963564653145639, + "accuracy": 0.9963640464763892, "total_bits": 784904192, "q_proj": { "group_size": { @@ -137475,7 +137475,7 @@ } }, { - "accuracy": 0.99722831029641, + "accuracy": 0.9972227968667683, "total_bits": 797982720, "q_proj": { "group_size": { @@ -137536,7 +137536,7 @@ } }, { - "accuracy": 0.9971540809461945, + "accuracy": 0.9971593314487683, "total_bits": 911912960, "q_proj": { "group_size": { @@ -137588,7 +137588,7 @@ } }, { - "accuracy": 0.99845988577918, + "accuracy": 0.9984637132208598, "total_bits": 942882816, "q_proj": { "group_size": { @@ -137640,7 +137640,7 @@ } }, { - "accuracy": 0.999248357587739, + "accuracy": 0.9992566953266138, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -137694,7 +137694,7 @@ ], "model.layers.70.mlp": [ { - "accuracy": 0.9364300464328966, + "accuracy": 0.9364626407623291, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -137746,7 +137746,7 @@ } }, { - "accuracy": 0.9380014507394088, + "accuracy": 0.9380452193711933, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -137798,7 +137798,7 @@ } }, { - "accuracy": 0.9462235664066515, + "accuracy": 0.9462868100718448, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -137847,7 +137847,7 @@ } }, { - "accuracy": 0.94879402612385, + "accuracy": 0.9488582705196581, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -137896,7 +137896,7 @@ } }, { - "accuracy": 0.9681140623594585, + "accuracy": 0.9681361882310164, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -137948,7 +137948,7 @@ } }, { - "accuracy": 0.9705998144651714, + "accuracy": 0.9706292717080367, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -138000,7 +138000,7 @@ } }, { - "accuracy": 0.9740886437265497, + "accuracy": 0.9741187440721613, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -138049,7 +138049,7 @@ } }, { - "accuracy": 0.9834545317449068, + "accuracy": 0.9834647492358559, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -138092,7 +138092,7 @@ } }, { - "accuracy": 0.984942050356614, + "accuracy": 0.9849528356602317, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -138135,7 +138135,7 @@ } }, { - "accuracy": 0.9838300551238813, + "accuracy": 0.9838466989366632, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -138187,7 +138187,7 @@ } }, { - "accuracy": 0.985814598045851, + "accuracy": 0.9858280624213972, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -138239,7 +138239,7 @@ } }, { - "accuracy": 0.9917863936800706, + "accuracy": 0.9917917416283959, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -138291,7 +138291,7 @@ } }, { - "accuracy": 0.9929353100688834, + "accuracy": 0.9929429951467013, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -138343,7 +138343,7 @@ } }, { - "accuracy": 0.995551201073747, + "accuracy": 0.9955541518957991, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -138386,7 +138386,7 @@ } }, { - "accuracy": 0.9958273376289167, + "accuracy": 0.9958347768375748, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -138435,7 +138435,7 @@ } }, { - "accuracy": 0.9965108872244233, + "accuracy": 0.9965179535119157, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -138481,7 +138481,7 @@ } }, { - "accuracy": 0.9987877987717327, + "accuracy": 0.9987891942850853, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -138523,7 +138523,7 @@ ], "model.layers.71.self_attn": [ { - "accuracy": 0.9724085205479672, + "accuracy": 0.9724369174555728, "total_bits": 320921600, "q_proj": { "group_size": { @@ -138587,7 +138587,7 @@ } }, { - "accuracy": 0.9732423016899511, + "accuracy": 0.9732936683454012, "total_bits": 329244672, "q_proj": { "group_size": { @@ -138651,7 +138651,7 @@ } }, { - "accuracy": 0.975536192718305, + "accuracy": 0.9755825337610746, "total_bits": 336188416, "q_proj": { "group_size": { @@ -138715,7 +138715,7 @@ } }, { - "accuracy": 0.9808367757420791, + "accuracy": 0.9808494134953147, "total_bits": 401721344, "q_proj": { "group_size": { @@ -138779,7 +138779,7 @@ } }, { - "accuracy": 0.9855938585180986, + "accuracy": 0.9856900415922466, "total_bits": 475443200, "q_proj": { "group_size": { @@ -138843,7 +138843,7 @@ } }, { - "accuracy": 0.9861499604425932, + "accuracy": 0.9861840919444436, "total_bits": 475642880, "q_proj": { "group_size": { @@ -138907,7 +138907,7 @@ } }, { - "accuracy": 0.9898735904379895, + "accuracy": 0.9899758273049405, "total_bits": 609923072, "q_proj": { "group_size": { @@ -138959,7 +138959,7 @@ } }, { - "accuracy": 0.9906393054284548, + "accuracy": 0.9906439059659055, "total_bits": 610188288, "q_proj": { "group_size": { @@ -139011,7 +139011,7 @@ } }, { - "accuracy": 0.9915389136264199, + "accuracy": 0.9915818697527835, "total_bits": 615184384, "q_proj": { "group_size": { @@ -139063,7 +139063,7 @@ } }, { - "accuracy": 0.9921005603514219, + "accuracy": 0.992194729415994, "total_bits": 624115712, "q_proj": { "group_size": { @@ -139115,7 +139115,7 @@ } }, { - "accuracy": 0.9928998468737853, + "accuracy": 0.992875379167105, "total_bits": 626637824, "q_proj": { "group_size": { @@ -139179,7 +139179,7 @@ } }, { - "accuracy": 0.9935188630693838, + "accuracy": 0.9935433209726685, "total_bits": 630519808, "q_proj": { "group_size": { @@ -139243,7 +139243,7 @@ } }, { - "accuracy": 0.9935693423214712, + "accuracy": 0.9935369440599492, "total_bits": 637526016, "q_proj": { "group_size": { @@ -139304,7 +139304,7 @@ } }, { - "accuracy": 0.994302618268289, + "accuracy": 0.9943261052432814, "total_bits": 646987776, "q_proj": { "group_size": { @@ -139365,7 +139365,7 @@ } }, { - "accuracy": 0.9963923139791739, + "accuracy": 0.9963780096486995, "total_bits": 784904192, "q_proj": { "group_size": { @@ -139426,7 +139426,7 @@ } }, { - "accuracy": 0.9971362834698275, + "accuracy": 0.99714618725212, "total_bits": 797982720, "q_proj": { "group_size": { @@ -139487,7 +139487,7 @@ } }, { - "accuracy": 0.9972233144860518, + "accuracy": 0.9972009190211171, "total_bits": 911912960, "q_proj": { "group_size": { @@ -139539,7 +139539,7 @@ } }, { - "accuracy": 0.9984056088877352, + "accuracy": 0.9984087048980751, "total_bits": 942882816, "q_proj": { "group_size": { @@ -139591,7 +139591,7 @@ } }, { - "accuracy": 0.9992047927195304, + "accuracy": 0.999213601658611, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -139645,7 +139645,7 @@ ], "model.layers.71.mlp": [ { - "accuracy": 0.9359050989151001, + "accuracy": 0.9359882442574752, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -139697,7 +139697,7 @@ } }, { - "accuracy": 0.9375088591324656, + "accuracy": 0.937543975679498, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -139749,7 +139749,7 @@ } }, { - "accuracy": 0.9456427285545751, + "accuracy": 0.9456661123978465, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -139798,7 +139798,7 @@ } }, { - "accuracy": 0.9482452932157015, + "accuracy": 0.9482676920137907, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -139847,7 +139847,7 @@ } }, { - "accuracy": 0.9678507102163214, + "accuracy": 0.9678748438232824, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -139899,7 +139899,7 @@ } }, { - "accuracy": 0.9703358537272403, + "accuracy": 0.9703662960152877, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -139951,7 +139951,7 @@ } }, { - "accuracy": 0.9738212321933947, + "accuracy": 0.9738503898444929, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -140000,7 +140000,7 @@ } }, { - "accuracy": 0.9832855541455118, + "accuracy": 0.9833074557153803, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -140043,7 +140043,7 @@ } }, { - "accuracy": 0.984791813712371, + "accuracy": 0.9848074913024902, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -140086,7 +140086,7 @@ } }, { - "accuracy": 0.9837073354344619, + "accuracy": 0.9837279209965154, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -140138,7 +140138,7 @@ } }, { - "accuracy": 0.985695010737369, + "accuracy": 0.9857080076870165, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -140190,7 +140190,7 @@ } }, { - "accuracy": 0.9917293630148235, + "accuracy": 0.9917401067520443, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -140242,7 +140242,7 @@ } }, { - "accuracy": 0.992876239513096, + "accuracy": 0.9928866664045736, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -140294,7 +140294,7 @@ } }, { - "accuracy": 0.9955183706785503, + "accuracy": 0.9955253193252965, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -140337,7 +140337,7 @@ } }, { - "accuracy": 0.9958041663232603, + "accuracy": 0.995808613535605, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -140386,7 +140386,7 @@ } }, { - "accuracy": 0.9964897138507742, + "accuracy": 0.9964936681483921, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -140432,7 +140432,7 @@ } }, { - "accuracy": 0.9987780357662001, + "accuracy": 0.9987795498026045, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -140474,7 +140474,7 @@ ], "model.layers.72.self_attn": [ { - "accuracy": 0.9719774691682113, + "accuracy": 0.9722424274996707, "total_bits": 320921600, "q_proj": { "group_size": { @@ -140538,7 +140538,7 @@ } }, { - "accuracy": 0.973020098711315, + "accuracy": 0.9733351061218664, "total_bits": 329244672, "q_proj": { "group_size": { @@ -140602,7 +140602,7 @@ } }, { - "accuracy": 0.9761179669907218, + "accuracy": 0.9762670005622663, "total_bits": 336188416, "q_proj": { "group_size": { @@ -140666,7 +140666,7 @@ } }, { - "accuracy": 0.9815987348556519, + "accuracy": 0.9817794357475481, "total_bits": 401721344, "q_proj": { "group_size": { @@ -140730,7 +140730,7 @@ } }, { - "accuracy": 0.9853950331085607, + "accuracy": 0.9853928622446562, "total_bits": 475443200, "q_proj": { "group_size": { @@ -140794,7 +140794,7 @@ } }, { - "accuracy": 0.9859887879145773, + "accuracy": 0.9860975161979073, "total_bits": 475642880, "q_proj": { "group_size": { @@ -140858,7 +140858,7 @@ } }, { - "accuracy": 0.9899996472032446, + "accuracy": 0.9899980127811432, "total_bits": 609923072, "q_proj": { "group_size": { @@ -140910,7 +140910,7 @@ } }, { - "accuracy": 0.9908207098120138, + "accuracy": 0.9909392101200003, "total_bits": 610188288, "q_proj": { "group_size": { @@ -140962,7 +140962,7 @@ } }, { - "accuracy": 0.9919837854410473, + "accuracy": 0.9920381291916496, "total_bits": 615184384, "q_proj": { "group_size": { @@ -141014,7 +141014,7 @@ } }, { - "accuracy": 0.992432461757409, + "accuracy": 0.9925081706360767, "total_bits": 624115712, "q_proj": { "group_size": { @@ -141066,7 +141066,7 @@ } }, { - "accuracy": 0.9928516502443113, + "accuracy": 0.9929067351316151, "total_bits": 626637824, "q_proj": { "group_size": { @@ -141130,7 +141130,7 @@ } }, { - "accuracy": 0.993508272265133, + "accuracy": 0.9935470678304371, "total_bits": 630519808, "q_proj": { "group_size": { @@ -141194,7 +141194,7 @@ } }, { - "accuracy": 0.9938240333607322, + "accuracy": 0.9938993128506761, "total_bits": 637526016, "q_proj": { "group_size": { @@ -141255,7 +141255,7 @@ } }, { - "accuracy": 0.994463326507493, + "accuracy": 0.9944999555223867, "total_bits": 646987776, "q_proj": { "group_size": { @@ -141316,7 +141316,7 @@ } }, { - "accuracy": 0.996433754108454, + "accuracy": 0.9964249914414004, "total_bits": 784904192, "q_proj": { "group_size": { @@ -141377,7 +141377,7 @@ } }, { - "accuracy": 0.997184823022077, + "accuracy": 0.9972048623389319, "total_bits": 797982720, "q_proj": { "group_size": { @@ -141438,7 +141438,7 @@ } }, { - "accuracy": 0.9972340307737652, + "accuracy": 0.9972177100808997, "total_bits": 911912960, "q_proj": { "group_size": { @@ -141490,7 +141490,7 @@ } }, { - "accuracy": 0.9984594738405002, + "accuracy": 0.9984688759830437, "total_bits": 942882816, "q_proj": { "group_size": { @@ -141542,7 +141542,7 @@ } }, { - "accuracy": 0.9992356059680644, + "accuracy": 0.999237338718223, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -141596,7 +141596,7 @@ ], "model.layers.72.mlp": [ { - "accuracy": 0.9353734254837036, + "accuracy": 0.9354754058938277, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -141648,7 +141648,7 @@ } }, { - "accuracy": 0.9369915472833734, + "accuracy": 0.9370602055599815, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -141700,7 +141700,7 @@ } }, { - "accuracy": 0.9447733477542275, + "accuracy": 0.9448474331905967, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -141749,7 +141749,7 @@ } }, { - "accuracy": 0.9473741430985301, + "accuracy": 0.9474452577139202, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -141798,7 +141798,7 @@ } }, { - "accuracy": 0.9675869722115367, + "accuracy": 0.9676058449243244, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -141850,7 +141850,7 @@ } }, { - "accuracy": 0.970103194839076, + "accuracy": 0.9701262806591234, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -141902,7 +141902,7 @@ } }, { - "accuracy": 0.9734785745018407, + "accuracy": 0.9735045997720015, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -141951,7 +141951,7 @@ } }, { - "accuracy": 0.9830621622110668, + "accuracy": 0.9830839979021173, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -141994,7 +141994,7 @@ } }, { - "accuracy": 0.984607396941436, + "accuracy": 0.9846232078577343, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -142037,7 +142037,7 @@ } }, { - "accuracy": 0.9835629635735562, + "accuracy": 0.9835872744259081, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -142089,7 +142089,7 @@ } }, { - "accuracy": 0.9855778954531017, + "accuracy": 0.9855908029957822, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -142141,7 +142141,7 @@ } }, { - "accuracy": 0.9916556155995319, + "accuracy": 0.9916659460255974, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -142193,7 +142193,7 @@ } }, { - "accuracy": 0.9928189225886997, + "accuracy": 0.9928228784548608, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -142245,7 +142245,7 @@ } }, { - "accuracy": 0.9954548046777123, + "accuracy": 0.995460622404751, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -142288,7 +142288,7 @@ } }, { - "accuracy": 0.9957705881250533, + "accuracy": 0.9957741326407382, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -142337,7 +142337,7 @@ } }, { - "accuracy": 0.9964354602914107, + "accuracy": 0.9964387581536645, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -142383,7 +142383,7 @@ } }, { - "accuracy": 0.9987525126259578, + "accuracy": 0.9987542672376883, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -142425,7 +142425,7 @@ ], "model.layers.73.self_attn": [ { - "accuracy": 0.9725118750020078, + "accuracy": 0.9726879910418862, "total_bits": 320921600, "q_proj": { "group_size": { @@ -142489,7 +142489,7 @@ } }, { - "accuracy": 0.9733755212081107, + "accuracy": 0.9735786318778992, "total_bits": 329244672, "q_proj": { "group_size": { @@ -142553,7 +142553,7 @@ } }, { - "accuracy": 0.9756352838717008, + "accuracy": 0.9758131911880091, "total_bits": 336188416, "q_proj": { "group_size": { @@ -142617,7 +142617,7 @@ } }, { - "accuracy": 0.9801285000223863, + "accuracy": 0.9802745235593695, "total_bits": 401721344, "q_proj": { "group_size": { @@ -142681,7 +142681,7 @@ } }, { - "accuracy": 0.9859258877603632, + "accuracy": 0.9859338559602436, "total_bits": 475443200, "q_proj": { "group_size": { @@ -142745,7 +142745,7 @@ } }, { - "accuracy": 0.9864662054337954, + "accuracy": 0.9864598073457417, "total_bits": 475642880, "q_proj": { "group_size": { @@ -142809,7 +142809,7 @@ } }, { - "accuracy": 0.9899361643352007, + "accuracy": 0.9899204036122874, "total_bits": 609923072, "q_proj": { "group_size": { @@ -142861,7 +142861,7 @@ } }, { - "accuracy": 0.9906422425257532, + "accuracy": 0.9906279570177982, "total_bits": 610188288, "q_proj": { "group_size": { @@ -142913,7 +142913,7 @@ } }, { - "accuracy": 0.9913462431807267, + "accuracy": 0.9913422033974999, "total_bits": 615184384, "q_proj": { "group_size": { @@ -142965,7 +142965,7 @@ } }, { - "accuracy": 0.9920696522060194, + "accuracy": 0.9920890888101176, "total_bits": 624115712, "q_proj": { "group_size": { @@ -143017,7 +143017,7 @@ } }, { - "accuracy": 0.9928464293479919, + "accuracy": 0.9928796950139498, "total_bits": 626637824, "q_proj": { "group_size": { @@ -143081,7 +143081,7 @@ } }, { - "accuracy": 0.9935600965430862, + "accuracy": 0.9935960193213663, "total_bits": 630519808, "q_proj": { "group_size": { @@ -143145,7 +143145,7 @@ } }, { - "accuracy": 0.99358660570885, + "accuracy": 0.9936289995124465, "total_bits": 637526016, "q_proj": { "group_size": { @@ -143206,7 +143206,7 @@ } }, { - "accuracy": 0.9942455170185942, + "accuracy": 0.9943089485168457, "total_bits": 646987776, "q_proj": { "group_size": { @@ -143267,7 +143267,7 @@ } }, { - "accuracy": 0.9963654969867907, + "accuracy": 0.9963721970194265, "total_bits": 784904192, "q_proj": { "group_size": { @@ -143328,7 +143328,7 @@ } }, { - "accuracy": 0.9971412612419379, + "accuracy": 0.9971360074062097, "total_bits": 797982720, "q_proj": { "group_size": { @@ -143389,7 +143389,7 @@ } }, { - "accuracy": 0.9971380878828073, + "accuracy": 0.9971401691436768, "total_bits": 911912960, "q_proj": { "group_size": { @@ -143441,7 +143441,7 @@ } }, { - "accuracy": 0.9984005763146439, + "accuracy": 0.998399897135402, "total_bits": 942882816, "q_proj": { "group_size": { @@ -143493,7 +143493,7 @@ } }, { - "accuracy": 0.9992043058829088, + "accuracy": 0.9992086553063831, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -143547,7 +143547,7 @@ ], "model.layers.73.mlp": [ { - "accuracy": 0.9359648980592427, + "accuracy": 0.9360492166719938, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -143599,7 +143599,7 @@ } }, { - "accuracy": 0.9376009765424227, + "accuracy": 0.9376466901678788, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -143651,7 +143651,7 @@ } }, { - "accuracy": 0.9449574821873715, + "accuracy": 0.9450388017453646, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -143700,7 +143700,7 @@ } }, { - "accuracy": 0.9474700563832333, + "accuracy": 0.9475492615448802, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -143749,7 +143749,7 @@ } }, { - "accuracy": 0.9678138431749845, + "accuracy": 0.9678498318320826, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -143801,7 +143801,7 @@ } }, { - "accuracy": 0.9703318413935209, + "accuracy": 0.9703685672659623, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -143853,7 +143853,7 @@ } }, { - "accuracy": 0.9735353996879176, + "accuracy": 0.9735754979284186, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -143902,7 +143902,7 @@ } }, { - "accuracy": 0.9831622901715731, + "accuracy": 0.9831772462317818, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -143945,7 +143945,7 @@ } }, { - "accuracy": 0.9846852798210947, + "accuracy": 0.9847000884382349, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -143988,7 +143988,7 @@ } }, { - "accuracy": 0.9836873819953517, + "accuracy": 0.983699597810444, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -144040,7 +144040,7 @@ } }, { - "accuracy": 0.9856962357696734, + "accuracy": 0.9857037632088912, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -144092,7 +144092,7 @@ } }, { - "accuracy": 0.9917187839746475, + "accuracy": 0.9917261647550684, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -144144,7 +144144,7 @@ } }, { - "accuracy": 0.9928767641908244, + "accuracy": 0.9928825387829229, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -144196,7 +144196,7 @@ } }, { - "accuracy": 0.9954796233459523, + "accuracy": 0.995485463032597, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -144239,7 +144239,7 @@ } }, { - "accuracy": 0.9957975607953573, + "accuracy": 0.9958018519376454, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -144288,7 +144288,7 @@ } }, { - "accuracy": 0.9964332270779108, + "accuracy": 0.9964377609522719, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -144334,7 +144334,7 @@ } }, { - "accuracy": 0.9987424645376833, + "accuracy": 0.998743787603943, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -144376,7 +144376,7 @@ ], "model.layers.74.self_attn": [ { - "accuracy": 0.9712982083621778, + "accuracy": 0.9713022018733778, "total_bits": 320921600, "q_proj": { "group_size": { @@ -144440,7 +144440,7 @@ } }, { - "accuracy": 0.9723012133648521, + "accuracy": 0.9722931761490671, "total_bits": 329244672, "q_proj": { "group_size": { @@ -144504,7 +144504,7 @@ } }, { - "accuracy": 0.9755630446107764, + "accuracy": 0.9755112654284427, "total_bits": 336188416, "q_proj": { "group_size": { @@ -144568,7 +144568,7 @@ } }, { - "accuracy": 0.9811696943483854, + "accuracy": 0.9811173551960995, "total_bits": 401721344, "q_proj": { "group_size": { @@ -144632,7 +144632,7 @@ } }, { - "accuracy": 0.9851402938365936, + "accuracy": 0.9851452551389995, "total_bits": 475443200, "q_proj": { "group_size": { @@ -144696,7 +144696,7 @@ } }, { - "accuracy": 0.9856584385821694, + "accuracy": 0.98567386363682, "total_bits": 475642880, "q_proj": { "group_size": { @@ -144760,7 +144760,7 @@ } }, { - "accuracy": 0.9902522995283729, + "accuracy": 0.9902145893950212, "total_bits": 609923072, "q_proj": { "group_size": { @@ -144812,7 +144812,7 @@ } }, { - "accuracy": 0.9909435165555853, + "accuracy": 0.9909337194342362, "total_bits": 610188288, "q_proj": { "group_size": { @@ -144864,7 +144864,7 @@ } }, { - "accuracy": 0.9915106437708202, + "accuracy": 0.9916476246557737, "total_bits": 615184384, "q_proj": { "group_size": { @@ -144916,7 +144916,7 @@ } }, { - "accuracy": 0.9921623805635854, + "accuracy": 0.9921873407928568, "total_bits": 624115712, "q_proj": { "group_size": { @@ -144968,7 +144968,7 @@ } }, { - "accuracy": 0.9927360446829545, + "accuracy": 0.9927581085970527, "total_bits": 626637824, "q_proj": { "group_size": { @@ -145032,7 +145032,7 @@ } }, { - "accuracy": 0.9933360571924009, + "accuracy": 0.9933103652376878, "total_bits": 630519808, "q_proj": { "group_size": { @@ -145096,7 +145096,7 @@ } }, { - "accuracy": 0.9937728831642553, + "accuracy": 0.9937477680413347, "total_bits": 637526016, "q_proj": { "group_size": { @@ -145157,7 +145157,7 @@ } }, { - "accuracy": 0.9943148360440606, + "accuracy": 0.9943315053456708, "total_bits": 646987776, "q_proj": { "group_size": { @@ -145218,7 +145218,7 @@ } }, { - "accuracy": 0.9964550674745911, + "accuracy": 0.996470704282585, "total_bits": 784904192, "q_proj": { "group_size": { @@ -145279,7 +145279,7 @@ } }, { - "accuracy": 0.9971627377365765, + "accuracy": 0.9971632775115339, "total_bits": 797982720, "q_proj": { "group_size": { @@ -145340,7 +145340,7 @@ } }, { - "accuracy": 0.9974051294358153, + "accuracy": 0.9974227884882375, "total_bits": 911912960, "q_proj": { "group_size": { @@ -145392,7 +145392,7 @@ } }, { - "accuracy": 0.9984582734147185, + "accuracy": 0.9984651943178553, "total_bits": 942882816, "q_proj": { "group_size": { @@ -145444,7 +145444,7 @@ } }, { - "accuracy": 0.9992994142108058, + "accuracy": 0.9992999100175343, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -145498,7 +145498,7 @@ ], "model.layers.74.mlp": [ { - "accuracy": 0.9354161839736135, + "accuracy": 0.9355023722899587, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -145550,7 +145550,7 @@ } }, { - "accuracy": 0.937048221889295, + "accuracy": 0.9371244781895688, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -145602,7 +145602,7 @@ } }, { - "accuracy": 0.9442197272652074, + "accuracy": 0.9442917924178273, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -145651,7 +145651,7 @@ } }, { - "accuracy": 0.9467546437916002, + "accuracy": 0.946825805463289, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -145700,7 +145700,7 @@ } }, { - "accuracy": 0.9674820052950006, + "accuracy": 0.9675233866039076, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -145752,7 +145752,7 @@ } }, { - "accuracy": 0.9700454191157692, + "accuracy": 0.9700732419365331, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -145804,7 +145804,7 @@ } }, { - "accuracy": 0.9731942572091755, + "accuracy": 0.9732217820067155, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -145853,7 +145853,7 @@ } }, { - "accuracy": 0.9829138642863223, + "accuracy": 0.982932514265964, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -145896,7 +145896,7 @@ } }, { - "accuracy": 0.9844687173241063, + "accuracy": 0.9844829534229479, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -145939,7 +145939,7 @@ } }, { - "accuracy": 0.9835195855090493, + "accuracy": 0.9835317872072521, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -145991,7 +145991,7 @@ } }, { - "accuracy": 0.9855376340840992, + "accuracy": 0.9855560864272871, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -146043,7 +146043,7 @@ } }, { - "accuracy": 0.9916216229137621, + "accuracy": 0.9916334685526396, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -146095,7 +146095,7 @@ } }, { - "accuracy": 0.9927938619726583, + "accuracy": 0.9928009996288701, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -146147,7 +146147,7 @@ } }, { - "accuracy": 0.9953976748021025, + "accuracy": 0.9954038470199233, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -146190,7 +146190,7 @@ } }, { - "accuracy": 0.995751576204049, + "accuracy": 0.9957585573980683, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -146239,7 +146239,7 @@ } }, { - "accuracy": 0.9963791786055816, + "accuracy": 0.996386895054265, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -146285,7 +146285,7 @@ } }, { - "accuracy": 0.9987029240123535, + "accuracy": 0.9987074492597267, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -146327,7 +146327,7 @@ ], "model.layers.75.self_attn": [ { - "accuracy": 0.9677504552038092, + "accuracy": 0.9676876883757741, "total_bits": 320921600, "q_proj": { "group_size": { @@ -146391,7 +146391,7 @@ } }, { - "accuracy": 0.96895180250469, + "accuracy": 0.9688966054665414, "total_bits": 329244672, "q_proj": { "group_size": { @@ -146455,7 +146455,7 @@ } }, { - "accuracy": 0.9727267026901245, + "accuracy": 0.9726249293277138, "total_bits": 336188416, "q_proj": { "group_size": { @@ -146519,7 +146519,7 @@ } }, { - "accuracy": 0.9790888491429781, + "accuracy": 0.9789564264448065, "total_bits": 401721344, "q_proj": { "group_size": { @@ -146583,7 +146583,7 @@ } }, { - "accuracy": 0.9835477659576818, + "accuracy": 0.9836340938743792, "total_bits": 475443200, "q_proj": { "group_size": { @@ -146647,7 +146647,7 @@ } }, { - "accuracy": 0.9840958149809587, + "accuracy": 0.9841808770832262, "total_bits": 475642880, "q_proj": { "group_size": { @@ -146711,7 +146711,7 @@ } }, { - "accuracy": 0.9892328310953943, + "accuracy": 0.9893507557479959, "total_bits": 609923072, "q_proj": { "group_size": { @@ -146763,7 +146763,7 @@ } }, { - "accuracy": 0.9899164650001024, + "accuracy": 0.9900562410291872, "total_bits": 610188288, "q_proj": { "group_size": { @@ -146815,7 +146815,7 @@ } }, { - "accuracy": 0.9907806343153903, + "accuracy": 0.9907532553923758, "total_bits": 615184384, "q_proj": { "group_size": { @@ -146867,7 +146867,7 @@ } }, { - "accuracy": 0.9911882830293555, + "accuracy": 0.9911610021402961, "total_bits": 624115712, "q_proj": { "group_size": { @@ -146919,7 +146919,7 @@ } }, { - "accuracy": 0.991845041513443, + "accuracy": 0.991904760661878, "total_bits": 626637824, "q_proj": { "group_size": { @@ -146983,7 +146983,7 @@ } }, { - "accuracy": 0.9923723536102396, + "accuracy": 0.9923694086702246, "total_bits": 630519808, "q_proj": { "group_size": { @@ -147047,7 +147047,7 @@ } }, { - "accuracy": 0.9929590985963219, + "accuracy": 0.9929645398729726, "total_bits": 637526016, "q_proj": { "group_size": { @@ -147108,7 +147108,7 @@ } }, { - "accuracy": 0.99362734117006, + "accuracy": 0.9936179197148273, "total_bits": 646987776, "q_proj": { "group_size": { @@ -147169,7 +147169,7 @@ } }, { - "accuracy": 0.9960330195332828, + "accuracy": 0.9960454651399663, "total_bits": 784904192, "q_proj": { "group_size": { @@ -147230,7 +147230,7 @@ } }, { - "accuracy": 0.9968015424122936, + "accuracy": 0.9967972346042332, "total_bits": 797982720, "q_proj": { "group_size": { @@ -147291,7 +147291,7 @@ } }, { - "accuracy": 0.9970540318049883, + "accuracy": 0.9970628044715053, "total_bits": 911912960, "q_proj": { "group_size": { @@ -147343,7 +147343,7 @@ } }, { - "accuracy": 0.9982368951957477, + "accuracy": 0.9982183448186046, "total_bits": 942882816, "q_proj": { "group_size": { @@ -147395,7 +147395,7 @@ } }, { - "accuracy": 0.9991639025117222, + "accuracy": 0.9991653870399061, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -147449,7 +147449,7 @@ ], "model.layers.75.mlp": [ { - "accuracy": 0.9333256859528392, + "accuracy": 0.9333978452180561, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -147501,7 +147501,7 @@ } }, { - "accuracy": 0.9350155905673379, + "accuracy": 0.9351054806458323, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -147553,7 +147553,7 @@ } }, { - "accuracy": 0.9423183453710455, + "accuracy": 0.9423741792377672, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -147602,7 +147602,7 @@ } }, { - "accuracy": 0.9449869582527562, + "accuracy": 0.945038870761269, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -147651,7 +147651,7 @@ } }, { - "accuracy": 0.9663854272742021, + "accuracy": 0.9664304727002194, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -147703,7 +147703,7 @@ } }, { - "accuracy": 0.9690354215471368, + "accuracy": 0.9690735308747542, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -147755,7 +147755,7 @@ } }, { - "accuracy": 0.9722607763190019, + "accuracy": 0.9722932232053656, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -147804,7 +147804,7 @@ } }, { - "accuracy": 0.9822312201324263, + "accuracy": 0.9822512143536618, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -147847,7 +147847,7 @@ } }, { - "accuracy": 0.9838454519447527, + "accuracy": 0.9838662069094809, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -147890,7 +147890,7 @@ } }, { - "accuracy": 0.9829367618811758, + "accuracy": 0.982956134959271, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -147942,7 +147942,7 @@ } }, { - "accuracy": 0.9850396686478665, + "accuracy": 0.985062903479526, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -147994,7 +147994,7 @@ } }, { - "accuracy": 0.9913196328439211, + "accuracy": 0.9913300220903597, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -148046,7 +148046,7 @@ } }, { - "accuracy": 0.9925346115702077, + "accuracy": 0.9925410786741659, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -148098,7 +148098,7 @@ } }, { - "accuracy": 0.9951773199595904, + "accuracy": 0.995183179253026, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -148141,7 +148141,7 @@ } }, { - "accuracy": 0.9956048087854135, + "accuracy": 0.9956107402318403, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -148190,7 +148190,7 @@ } }, { - "accuracy": 0.996252218359395, + "accuracy": 0.9962574149432936, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -148236,7 +148236,7 @@ } }, { - "accuracy": 0.9986234887650138, + "accuracy": 0.9986261260744772, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -148278,7 +148278,7 @@ ], "model.layers.76.self_attn": [ { - "accuracy": 0.9654615521430969, + "accuracy": 0.9656241191060919, "total_bits": 320921600, "q_proj": { "group_size": { @@ -148342,7 +148342,7 @@ } }, { - "accuracy": 0.9666018140943426, + "accuracy": 0.9667643089043467, "total_bits": 329244672, "q_proj": { "group_size": { @@ -148406,7 +148406,7 @@ } }, { - "accuracy": 0.9708484756319147, + "accuracy": 0.9709304320184808, "total_bits": 336188416, "q_proj": { "group_size": { @@ -148470,7 +148470,7 @@ } }, { - "accuracy": 0.9769810046020307, + "accuracy": 0.9770420821089494, "total_bits": 401721344, "q_proj": { "group_size": { @@ -148534,7 +148534,7 @@ } }, { - "accuracy": 0.9822614255704378, + "accuracy": 0.9820876435229653, "total_bits": 475443200, "q_proj": { "group_size": { @@ -148598,7 +148598,7 @@ } }, { - "accuracy": 0.9826274906334124, + "accuracy": 0.982441213570143, "total_bits": 475642880, "q_proj": { "group_size": { @@ -148662,7 +148662,7 @@ } }, { - "accuracy": 0.9883067631407788, + "accuracy": 0.9880125404972779, "total_bits": 609923072, "q_proj": { "group_size": { @@ -148714,7 +148714,7 @@ } }, { - "accuracy": 0.9887177928497917, + "accuracy": 0.9884064730844999, "total_bits": 610188288, "q_proj": { "group_size": { @@ -148766,7 +148766,7 @@ } }, { - "accuracy": 0.9895866121116438, + "accuracy": 0.9896411785953924, "total_bits": 615184384, "q_proj": { "group_size": { @@ -148818,7 +148818,7 @@ } }, { - "accuracy": 0.9905125145849428, + "accuracy": 0.9905683296291452, "total_bits": 624115712, "q_proj": { "group_size": { @@ -148870,7 +148870,7 @@ } }, { - "accuracy": 0.9910211641537515, + "accuracy": 0.9910838407905478, "total_bits": 626637824, "q_proj": { "group_size": { @@ -148934,7 +148934,7 @@ } }, { - "accuracy": 0.9919093298284631, + "accuracy": 0.9919111211048929, "total_bits": 630519808, "q_proj": { "group_size": { @@ -148998,7 +148998,7 @@ } }, { - "accuracy": 0.9924409005202746, + "accuracy": 0.9924593442364743, "total_bits": 637526016, "q_proj": { "group_size": { @@ -149059,7 +149059,7 @@ } }, { - "accuracy": 0.9932268677573455, + "accuracy": 0.9932717594661211, "total_bits": 646987776, "q_proj": { "group_size": { @@ -149120,7 +149120,7 @@ } }, { - "accuracy": 0.9957658738682145, + "accuracy": 0.9957708810505114, "total_bits": 784904192, "q_proj": { "group_size": { @@ -149181,7 +149181,7 @@ } }, { - "accuracy": 0.9966039547794744, + "accuracy": 0.996643200124565, "total_bits": 797982720, "q_proj": { "group_size": { @@ -149242,7 +149242,7 @@ } }, { - "accuracy": 0.9967878223641923, + "accuracy": 0.9967776697717214, "total_bits": 911912960, "q_proj": { "group_size": { @@ -149294,7 +149294,7 @@ } }, { - "accuracy": 0.9981103986501694, + "accuracy": 0.99811798118447, "total_bits": 942882816, "q_proj": { "group_size": { @@ -149346,7 +149346,7 @@ } }, { - "accuracy": 0.9991206012078022, + "accuracy": 0.9991233691965279, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -149400,7 +149400,7 @@ ], "model.layers.76.mlp": [ { - "accuracy": 0.9318119663941233, + "accuracy": 0.9319115249734176, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -149452,7 +149452,7 @@ } }, { - "accuracy": 0.9335895212073075, + "accuracy": 0.9336407812018144, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -149504,7 +149504,7 @@ } }, { - "accuracy": 0.9411837239014476, + "accuracy": 0.9412563976488615, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -149553,7 +149553,7 @@ } }, { - "accuracy": 0.9439946726748818, + "accuracy": 0.9440657465081466, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -149602,7 +149602,7 @@ } }, { - "accuracy": 0.9654434009602195, + "accuracy": 0.9654672647777357, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -149654,7 +149654,7 @@ } }, { - "accuracy": 0.9682105371826574, + "accuracy": 0.9682445682977375, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -149706,7 +149706,7 @@ } }, { - "accuracy": 0.9716548637339943, + "accuracy": 0.9716891928722984, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -149755,7 +149755,7 @@ } }, { - "accuracy": 0.9816251381447441, + "accuracy": 0.9816392801309887, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -149798,7 +149798,7 @@ } }, { - "accuracy": 0.9832929498270938, + "accuracy": 0.9833119370435414, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -149841,7 +149841,7 @@ } }, { - "accuracy": 0.982372931743923, + "accuracy": 0.982383121001093, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -149893,7 +149893,7 @@ } }, { - "accuracy": 0.9845647780518783, + "accuracy": 0.9845801560502303, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -149945,7 +149945,7 @@ } }, { - "accuracy": 0.9909992790535876, + "accuracy": 0.9910031274745339, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -149997,7 +149997,7 @@ } }, { - "accuracy": 0.9922886984912973, + "accuracy": 0.9922951114805121, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -150049,7 +150049,7 @@ } }, { - "accuracy": 0.9949850667464105, + "accuracy": 0.9949878524792822, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -150092,7 +150092,7 @@ } }, { - "accuracy": 0.9954337120840424, + "accuracy": 0.9954356119820946, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -150141,7 +150141,7 @@ } }, { - "accuracy": 0.9961419944700441, + "accuracy": 0.996143739866583, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -150187,7 +150187,7 @@ } }, { - "accuracy": 0.9985897517517993, + "accuracy": 0.9985903709342605, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -150229,7 +150229,7 @@ ], "model.layers.77.self_attn": [ { - "accuracy": 0.966634728406605, + "accuracy": 0.9667535486974215, "total_bits": 320921600, "q_proj": { "group_size": { @@ -150293,7 +150293,7 @@ } }, { - "accuracy": 0.9676372816688136, + "accuracy": 0.9678019975361071, "total_bits": 329244672, "q_proj": { "group_size": { @@ -150357,7 +150357,7 @@ } }, { - "accuracy": 0.9716581043444181, + "accuracy": 0.9718163201683446, "total_bits": 336188416, "q_proj": { "group_size": { @@ -150421,7 +150421,7 @@ } }, { - "accuracy": 0.978279140434767, + "accuracy": 0.9784191752734938, "total_bits": 401721344, "q_proj": { "group_size": { @@ -150485,7 +150485,7 @@ } }, { - "accuracy": 0.9831264191552213, + "accuracy": 0.9832410043791721, "total_bits": 475443200, "q_proj": { "group_size": { @@ -150549,7 +150549,7 @@ } }, { - "accuracy": 0.9837962969353324, + "accuracy": 0.9838592613998213, "total_bits": 475642880, "q_proj": { "group_size": { @@ -150613,7 +150613,7 @@ } }, { - "accuracy": 0.9891103960965809, + "accuracy": 0.9892157857355318, "total_bits": 609923072, "q_proj": { "group_size": { @@ -150665,7 +150665,7 @@ } }, { - "accuracy": 0.9900173521355579, + "accuracy": 0.990079895446175, "total_bits": 610188288, "q_proj": { "group_size": { @@ -150717,7 +150717,7 @@ } }, { - "accuracy": 0.9905879623011539, + "accuracy": 0.9906422927191383, "total_bits": 615184384, "q_proj": { "group_size": { @@ -150769,7 +150769,7 @@ } }, { - "accuracy": 0.9907047387800718, + "accuracy": 0.990751702534525, "total_bits": 624115712, "q_proj": { "group_size": { @@ -150821,7 +150821,7 @@ } }, { - "accuracy": 0.9917991271144465, + "accuracy": 0.9918143843349657, "total_bits": 626637824, "q_proj": { "group_size": { @@ -150885,7 +150885,7 @@ } }, { - "accuracy": 0.9923179502549925, + "accuracy": 0.9923509884821741, "total_bits": 630519808, "q_proj": { "group_size": { @@ -150949,7 +150949,7 @@ } }, { - "accuracy": 0.9929451997342863, + "accuracy": 0.9929695858767158, "total_bits": 637526016, "q_proj": { "group_size": { @@ -151010,7 +151010,7 @@ } }, { - "accuracy": 0.9934891658393961, + "accuracy": 0.9935012110754063, "total_bits": 646987776, "q_proj": { "group_size": { @@ -151071,7 +151071,7 @@ } }, { - "accuracy": 0.9959845593885371, + "accuracy": 0.9959971210674236, "total_bits": 784904192, "q_proj": { "group_size": { @@ -151132,7 +151132,7 @@ } }, { - "accuracy": 0.9967195066182237, + "accuracy": 0.9967233024929699, "total_bits": 797982720, "q_proj": { "group_size": { @@ -151193,7 +151193,7 @@ } }, { - "accuracy": 0.99709746947414, + "accuracy": 0.9971099427263987, "total_bits": 911912960, "q_proj": { "group_size": { @@ -151245,7 +151245,7 @@ } }, { - "accuracy": 0.9982195686745016, + "accuracy": 0.9982363430685118, "total_bits": 942882816, "q_proj": { "group_size": { @@ -151297,7 +151297,7 @@ } }, { - "accuracy": 0.9992142259388378, + "accuracy": 0.9992016173997208, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -151351,7 +151351,7 @@ ], "model.layers.77.mlp": [ { - "accuracy": 0.9263010213249608, + "accuracy": 0.9263562779677541, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -151403,7 +151403,7 @@ } }, { - "accuracy": 0.9281314674176668, + "accuracy": 0.9282026604602212, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -151455,7 +151455,7 @@ } }, { - "accuracy": 0.9364000307886224, + "accuracy": 0.9364434480667114, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -151504,7 +151504,7 @@ } }, { - "accuracy": 0.9394501447677612, + "accuracy": 0.9394889944478085, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -151553,7 +151553,7 @@ } }, { - "accuracy": 0.9624753029722917, + "accuracy": 0.9625128005680285, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -151605,7 +151605,7 @@ } }, { - "accuracy": 0.9655887139470953, + "accuracy": 0.9656107613914892, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -151657,7 +151657,7 @@ } }, { - "accuracy": 0.9692829910077547, + "accuracy": 0.9693082131837544, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -151706,7 +151706,7 @@ } }, { - "accuracy": 0.9799046610531054, + "accuracy": 0.9799362154383409, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -151749,7 +151749,7 @@ } }, { - "accuracy": 0.9817840958896437, + "accuracy": 0.9818043238238284, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -151792,7 +151792,7 @@ } }, { - "accuracy": 0.9807986287694228, + "accuracy": 0.9808201131067777, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -151844,7 +151844,7 @@ } }, { - "accuracy": 0.9831977549352144, + "accuracy": 0.9832060070414292, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -151896,7 +151896,7 @@ } }, { - "accuracy": 0.9901539063767383, + "accuracy": 0.9901675966225172, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -151948,7 +151948,7 @@ } }, { - "accuracy": 0.9915796698708284, + "accuracy": 0.9915874521983298, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -152000,7 +152000,7 @@ } }, { - "accuracy": 0.9944877361780718, + "accuracy": 0.9944973038999658, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -152043,7 +152043,7 @@ } }, { - "accuracy": 0.9950003545535239, + "accuracy": 0.9950070345872327, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -152092,7 +152092,7 @@ } }, { - "accuracy": 0.9958167127088496, + "accuracy": 0.9958240346688974, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -152138,7 +152138,7 @@ } }, { - "accuracy": 0.9984520890406872, + "accuracy": 0.9984549091834771, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -152180,7 +152180,7 @@ ], "model.layers.78.self_attn": [ { - "accuracy": 0.9676422853218882, + "accuracy": 0.9677692651748657, "total_bits": 320921600, "q_proj": { "group_size": { @@ -152244,7 +152244,7 @@ } }, { - "accuracy": 0.96896315248389, + "accuracy": 0.9689417763760215, "total_bits": 329244672, "q_proj": { "group_size": { @@ -152308,7 +152308,7 @@ } }, { - "accuracy": 0.9729341017572504, + "accuracy": 0.9729377627372742, "total_bits": 336188416, "q_proj": { "group_size": { @@ -152372,7 +152372,7 @@ } }, { - "accuracy": 0.980330322918139, + "accuracy": 0.980319763484754, "total_bits": 401721344, "q_proj": { "group_size": { @@ -152436,7 +152436,7 @@ } }, { - "accuracy": 0.9839005140881789, + "accuracy": 0.9839398813875098, "total_bits": 475443200, "q_proj": { "group_size": { @@ -152500,7 +152500,7 @@ } }, { - "accuracy": 0.9843582307037554, + "accuracy": 0.9843455066806391, "total_bits": 475642880, "q_proj": { "group_size": { @@ -152564,7 +152564,7 @@ } }, { - "accuracy": 0.989999068410773, + "accuracy": 0.9899817085579822, "total_bits": 609923072, "q_proj": { "group_size": { @@ -152616,7 +152616,7 @@ } }, { - "accuracy": 0.9905751574980585, + "accuracy": 0.9906016031378194, "total_bits": 610188288, "q_proj": { "group_size": { @@ -152668,7 +152668,7 @@ } }, { - "accuracy": 0.9911835295589346, + "accuracy": 0.9911489753346694, "total_bits": 615184384, "q_proj": { "group_size": { @@ -152720,7 +152720,7 @@ } }, { - "accuracy": 0.9914620157919432, + "accuracy": 0.9914862231204384, "total_bits": 624115712, "q_proj": { "group_size": { @@ -152772,7 +152772,7 @@ } }, { - "accuracy": 0.992058044985721, + "accuracy": 0.9920747452660611, "total_bits": 626637824, "q_proj": { "group_size": { @@ -152836,7 +152836,7 @@ } }, { - "accuracy": 0.9925685893548163, + "accuracy": 0.9925746305992729, "total_bits": 630519808, "q_proj": { "group_size": { @@ -152900,7 +152900,7 @@ } }, { - "accuracy": 0.9932113956463965, + "accuracy": 0.9932198453890649, "total_bits": 637526016, "q_proj": { "group_size": { @@ -152961,7 +152961,7 @@ } }, { - "accuracy": 0.9937515376429809, + "accuracy": 0.9937666983980882, "total_bits": 646987776, "q_proj": { "group_size": { @@ -153022,7 +153022,7 @@ } }, { - "accuracy": 0.9961693122198707, + "accuracy": 0.9961762040069229, "total_bits": 784904192, "q_proj": { "group_size": { @@ -153083,7 +153083,7 @@ } }, { - "accuracy": 0.9967720786991873, + "accuracy": 0.9967829698794767, "total_bits": 797982720, "q_proj": { "group_size": { @@ -153144,7 +153144,7 @@ } }, { - "accuracy": 0.9972091401486021, + "accuracy": 0.9972130973872385, "total_bits": 911912960, "q_proj": { "group_size": { @@ -153196,7 +153196,7 @@ } }, { - "accuracy": 0.9981697733072858, + "accuracy": 0.9981972877132265, "total_bits": 942882816, "q_proj": { "group_size": { @@ -153248,7 +153248,7 @@ } }, { - "accuracy": 0.9991891560585875, + "accuracy": 0.9991878834798148, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -153302,7 +153302,7 @@ ], "model.layers.78.mlp": [ { - "accuracy": 0.9285163314718949, + "accuracy": 0.9284080204210783, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -153354,7 +153354,7 @@ } }, { - "accuracy": 0.9304605283235249, + "accuracy": 0.9304760443536859, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -153406,7 +153406,7 @@ } }, { - "accuracy": 0.9380497305016768, + "accuracy": 0.9380748711134258, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -153455,7 +153455,7 @@ } }, { - "accuracy": 0.9410336770509419, + "accuracy": 0.941052606231288, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -153504,7 +153504,7 @@ } }, { - "accuracy": 0.9640504153151261, + "accuracy": 0.9640432407981471, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -153556,7 +153556,7 @@ } }, { - "accuracy": 0.9670586868336326, + "accuracy": 0.9671047392644381, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -153608,7 +153608,7 @@ } }, { - "accuracy": 0.9705999838678461, + "accuracy": 0.9706489506520724, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -153657,7 +153657,7 @@ } }, { - "accuracy": 0.9805408010357305, + "accuracy": 0.9805256881211933, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -153700,7 +153700,7 @@ } }, { - "accuracy": 0.9824367855724535, + "accuracy": 0.9824428166213789, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -153743,7 +153743,7 @@ } }, { - "accuracy": 0.9816027879714966, + "accuracy": 0.981611618870183, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -153795,7 +153795,7 @@ } }, { - "accuracy": 0.9839587886082498, + "accuracy": 0.9839608088919991, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -153847,7 +153847,7 @@ } }, { - "accuracy": 0.9905405515118649, + "accuracy": 0.9905419443783007, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -153899,7 +153899,7 @@ } }, { - "accuracy": 0.9919380749526777, + "accuracy": 0.9919432056577582, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -153951,7 +153951,7 @@ } }, { - "accuracy": 0.9946202816147554, + "accuracy": 0.9946199294767881, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -153994,7 +153994,7 @@ } }, { - "accuracy": 0.9951981188435304, + "accuracy": 0.9951999089435527, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -154043,7 +154043,7 @@ } }, { - "accuracy": 0.9960384870830336, + "accuracy": 0.9960408528384409, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -154089,7 +154089,7 @@ } }, { - "accuracy": 0.9984478547580933, + "accuracy": 0.9984432492209109, "total_bits": 5837501264, "gate_proj": { "group_size": { @@ -154131,7 +154131,7 @@ ], "model.layers.79.self_attn": [ { - "accuracy": 0.9875766799638146, + "accuracy": 0.9875980890110919, "total_bits": 320921600, "q_proj": { "group_size": { @@ -154195,7 +154195,7 @@ } }, { - "accuracy": 0.9881629763465178, + "accuracy": 0.9881700481239118, "total_bits": 329244672, "q_proj": { "group_size": { @@ -154259,7 +154259,7 @@ } }, { - "accuracy": 0.9900833870235243, + "accuracy": 0.9900941707585987, "total_bits": 336188416, "q_proj": { "group_size": { @@ -154323,7 +154323,7 @@ } }, { - "accuracy": 0.9921991276113611, + "accuracy": 0.9922084518169102, "total_bits": 401721344, "q_proj": { "group_size": { @@ -154387,7 +154387,7 @@ } }, { - "accuracy": 0.993973013601805, + "accuracy": 0.993989893480351, "total_bits": 475443200, "q_proj": { "group_size": { @@ -154451,7 +154451,7 @@ } }, { - "accuracy": 0.9940683755435442, + "accuracy": 0.9940837594239336, "total_bits": 475642880, "q_proj": { "group_size": { @@ -154515,7 +154515,7 @@ } }, { - "accuracy": 0.996610052883625, + "accuracy": 0.9966219510687025, "total_bits": 609923072, "q_proj": { "group_size": { @@ -154567,7 +154567,7 @@ } }, { - "accuracy": 0.9966996119994866, + "accuracy": 0.9967062461533045, "total_bits": 610188288, "q_proj": { "group_size": { @@ -154619,7 +154619,7 @@ } }, { - "accuracy": 0.9969029016792774, + "accuracy": 0.9969222671107242, "total_bits": 615184384, "q_proj": { "group_size": { @@ -154671,7 +154671,7 @@ } }, { - "accuracy": 0.9970434176686563, + "accuracy": 0.9970524401256913, "total_bits": 624115712, "q_proj": { "group_size": { @@ -154723,7 +154723,7 @@ } }, { - "accuracy": 0.9969735471041579, + "accuracy": 0.996978256263231, "total_bits": 626637824, "q_proj": { "group_size": { @@ -154787,7 +154787,7 @@ } }, { - "accuracy": 0.9971509473888498, + "accuracy": 0.9971620269904011, "total_bits": 630519808, "q_proj": { "group_size": { @@ -154851,7 +154851,7 @@ } }, { - "accuracy": 0.997527071323834, + "accuracy": 0.9975349363919935, "total_bits": 637526016, "q_proj": { "group_size": { @@ -154912,7 +154912,7 @@ } }, { - "accuracy": 0.9977026548432676, + "accuracy": 0.9977080265158101, "total_bits": 646987776, "q_proj": { "group_size": { @@ -154973,7 +154973,7 @@ } }, { - "accuracy": 0.9986250785816657, + "accuracy": 0.9986310807105742, "total_bits": 784904192, "q_proj": { "group_size": { @@ -155034,7 +155034,7 @@ } }, { - "accuracy": 0.9988073301746657, + "accuracy": 0.9988148336935985, "total_bits": 797982720, "q_proj": { "group_size": { @@ -155095,7 +155095,7 @@ } }, { - "accuracy": 0.9990647500871044, + "accuracy": 0.9990707446673983, "total_bits": 911912960, "q_proj": { "group_size": { @@ -155147,7 +155147,7 @@ } }, { - "accuracy": 0.9993445273292693, + "accuracy": 0.9993481612029044, "total_bits": 942882816, "q_proj": { "group_size": { @@ -155199,7 +155199,7 @@ } }, { - "accuracy": 0.9996690078099307, + "accuracy": 0.9996712402636675, "total_bits": 1213902848, "q_proj": { "group_size": { @@ -155253,7 +155253,7 @@ ], "model.layers.79.mlp": [ { - "accuracy": 0.9320744777980604, + "accuracy": 0.9318859263470298, "total_bits": 1625570080, "gate_proj": { "group_size": { @@ -155305,7 +155305,7 @@ } }, { - "accuracy": 0.9345066296426874, + "accuracy": 0.9345646155507941, "total_bits": 1684232992, "gate_proj": { "group_size": { @@ -155357,7 +155357,7 @@ } }, { - "accuracy": 0.9396917129817762, + "accuracy": 0.9396193843138845, "total_bits": 1880673600, "gate_proj": { "group_size": { @@ -155406,7 +155406,7 @@ } }, { - "accuracy": 0.9426769331881875, + "accuracy": 0.942632330091376, "total_bits": 2110573888, "gate_proj": { "group_size": { @@ -155455,7 +155455,7 @@ } }, { - "accuracy": 0.9667846466365614, + "accuracy": 0.9666988473189504, "total_bits": 2379184656, "gate_proj": { "group_size": { @@ -155507,7 +155507,7 @@ } }, { - "accuracy": 0.9694773586172807, + "accuracy": 0.9695139935142115, "total_bits": 2443500864, "gate_proj": { "group_size": { @@ -155559,7 +155559,7 @@ } }, { - "accuracy": 0.9719593838641518, + "accuracy": 0.9719857385284022, "total_bits": 2627461648, "gate_proj": { "group_size": { @@ -155608,7 +155608,7 @@ } }, { - "accuracy": 0.9812871390267423, + "accuracy": 0.9812515221144024, "total_bits": 3004030480, "gate_proj": { "group_size": { @@ -155651,7 +155651,7 @@ } }, { - "accuracy": 0.9830765645754965, + "accuracy": 0.9830687124478189, "total_bits": 3048308032, "gate_proj": { "group_size": { @@ -155694,7 +155694,7 @@ } }, { - "accuracy": 0.983030013347927, + "accuracy": 0.9830465944189775, "total_bits": 3093527056, "gate_proj": { "group_size": { @@ -155746,7 +155746,7 @@ } }, { - "accuracy": 0.9852288867297926, + "accuracy": 0.9852246971506822, "total_bits": 3157843264, "gate_proj": { "group_size": { @@ -155798,7 +155798,7 @@ } }, { - "accuracy": 0.9911629957588095, + "accuracy": 0.9911833679989764, "total_bits": 3807869456, "gate_proj": { "group_size": { @@ -155850,7 +155850,7 @@ } }, { - "accuracy": 0.9924280094472986, + "accuracy": 0.9924506819561908, "total_bits": 3872185664, "gate_proj": { "group_size": { @@ -155902,7 +155902,7 @@ } }, { - "accuracy": 0.9945351324583355, + "accuracy": 0.9945806970721797, "total_bits": 4409997840, "gate_proj": { "group_size": { @@ -155945,7 +155945,7 @@ } }, { - "accuracy": 0.9954485944227168, + "accuracy": 0.9954518526792526, "total_bits": 4563546960, "gate_proj": { "group_size": { @@ -155994,7 +155994,7 @@ } }, { - "accuracy": 0.9961172295244116, + "accuracy": 0.9961240844506967, "total_bits": 4974588752, "gate_proj": { "group_size": { @@ -156040,7 +156040,7 @@ } }, { - "accuracy": 0.9980961682372972, + "accuracy": 0.9980900177830144, "total_bits": 5837501264, "gate_proj": { "group_size": {