{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 385,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": -1.7278180122375488,
      "logits/rejected": -1.7377450466156006,
      "logps/chosen": -29.553977966308594,
      "logps/rejected": -42.813133239746094,
      "loss": 1.0,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": -1.8664803504943848,
      "logits/rejected": -1.8707994222640991,
      "logps/chosen": -36.978511810302734,
      "logps/rejected": -33.66939163208008,
      "loss": 0.9993,
      "rewards/accuracies": 0.5694444179534912,
      "rewards/chosen": 0.00028087408281862736,
      "rewards/margins": 0.0006740752141922712,
      "rewards/rejected": -0.00039320107316598296,
      "step": 10
    },
    {
      "epoch": 0.05,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": -1.9984451532363892,
      "logits/rejected": -2.0010995864868164,
      "logps/chosen": -29.63176918029785,
      "logps/rejected": -29.05954933166504,
      "loss": 0.9999,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": 0.00010425634536659345,
      "rewards/margins": 6.528960511786863e-05,
      "rewards/rejected": 3.8966707506915554e-05,
      "step": 20
    },
    {
      "epoch": 0.08,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": -1.9210799932479858,
      "logits/rejected": -1.9183847904205322,
      "logps/chosen": -31.414783477783203,
      "logps/rejected": -33.19659423828125,
      "loss": 1.0002,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": 1.3138540452928282e-05,
      "rewards/margins": -0.00020548875909298658,
      "rewards/rejected": 0.00021862727589905262,
      "step": 30
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": -2.0177221298217773,
      "logits/rejected": -2.008965492248535,
      "logps/chosen": -32.57322311401367,
      "logps/rejected": -32.500308990478516,
      "loss": 1.0001,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": 3.584356090868823e-05,
      "rewards/margins": -8.777440234553069e-05,
      "rewards/rejected": 0.00012361796689219773,
      "step": 40
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": -1.8622690439224243,
      "logits/rejected": -1.851509690284729,
      "logps/chosen": -33.547603607177734,
      "logps/rejected": -35.463592529296875,
      "loss": 0.9998,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 9.876764670480043e-05,
      "rewards/margins": 0.0002112251240760088,
      "rewards/rejected": -0.00011245747737120837,
      "step": 50
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": -1.9400131702423096,
      "logits/rejected": -1.9419806003570557,
      "logps/chosen": -32.52842330932617,
      "logps/rejected": -33.22877883911133,
      "loss": 0.9987,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.0007175664068199694,
      "rewards/margins": 0.0012915965635329485,
      "rewards/rejected": -0.00057403021492064,
      "step": 60
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": -2.070552349090576,
      "logits/rejected": -2.0755274295806885,
      "logps/chosen": -34.00461959838867,
      "logps/rejected": -36.64922332763672,
      "loss": 0.9994,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": -0.00016131921438500285,
      "rewards/margins": 0.0006422021542675793,
      "rewards/rejected": -0.0008035213686525822,
      "step": 70
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": -1.9306777715682983,
      "logits/rejected": -1.9338254928588867,
      "logps/chosen": -34.32624816894531,
      "logps/rejected": -34.661468505859375,
      "loss": 0.9983,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.0010156143689528108,
      "rewards/margins": 0.0017062196275219321,
      "rewards/rejected": -0.0006906053749844432,
      "step": 80
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": -1.9389193058013916,
      "logits/rejected": -1.9434226751327515,
      "logps/chosen": -32.38957214355469,
      "logps/rejected": -32.348140716552734,
      "loss": 0.9993,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.0007785108755342662,
      "rewards/margins": 0.0006708315922878683,
      "rewards/rejected": 0.00010767912317533046,
      "step": 90
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": -2.0358777046203613,
      "logits/rejected": -2.0339014530181885,
      "logps/chosen": -32.13254165649414,
      "logps/rejected": -31.29019546508789,
      "loss": 0.9987,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.0010829826351255178,
      "rewards/margins": 0.001322855008766055,
      "rewards/rejected": -0.0002398724900558591,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": -2.2312774658203125,
      "eval_logits/rejected": -2.226422071456909,
      "eval_logps/chosen": -34.04991149902344,
      "eval_logps/rejected": -37.55283737182617,
      "eval_loss": 0.9997907280921936,
      "eval_rewards/accuracies": 0.5336378812789917,
      "eval_rewards/chosen": -0.0001535558985779062,
      "eval_rewards/margins": 0.00020861340453848243,
      "eval_rewards/rejected": -0.00036216925946064293,
      "eval_runtime": 146.0254,
      "eval_samples_per_second": 2.349,
      "eval_steps_per_second": 0.294,
      "step": 100
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": -1.9907060861587524,
      "logits/rejected": -1.9883339405059814,
      "logps/chosen": -33.13169860839844,
      "logps/rejected": -34.033958435058594,
      "loss": 0.9988,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.0011211589444428682,
      "rewards/margins": 0.0011670273961499333,
      "rewards/rejected": -4.586850991472602e-05,
      "step": 110
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": -2.002023458480835,
      "logits/rejected": -1.993699312210083,
      "logps/chosen": -32.341697692871094,
      "logps/rejected": -32.16511917114258,
      "loss": 0.9989,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.0010320657165721059,
      "rewards/margins": 0.001068194629624486,
      "rewards/rejected": -3.612901855376549e-05,
      "step": 120
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": -2.028505802154541,
      "logits/rejected": -2.020526885986328,
      "logps/chosen": -30.3519287109375,
      "logps/rejected": -32.101314544677734,
      "loss": 0.9983,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.0012378758983686566,
      "rewards/margins": 0.001720982021652162,
      "rewards/rejected": -0.0004831062688026577,
      "step": 130
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": -1.9588673114776611,
      "logits/rejected": -1.9690834283828735,
      "logps/chosen": -31.205490112304688,
      "logps/rejected": -32.55961608886719,
      "loss": 0.9976,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.0021417266689240932,
      "rewards/margins": 0.0023890691809356213,
      "rewards/rejected": -0.00024734257021918893,
      "step": 140
    },
    {
      "epoch": 0.39,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": -1.8695415258407593,
      "logits/rejected": -1.8707062005996704,
      "logps/chosen": -33.88127899169922,
      "logps/rejected": -34.7686653137207,
      "loss": 0.9968,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.0031141345389187336,
      "rewards/margins": 0.003228238318115473,
      "rewards/rejected": -0.00011410393926780671,
      "step": 150
    },
    {
      "epoch": 0.42,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": -1.9212032556533813,
      "logits/rejected": -1.9178003072738647,
      "logps/chosen": -35.99773406982422,
      "logps/rejected": -32.705848693847656,
      "loss": 0.9984,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.0017394202295690775,
      "rewards/margins": 0.0015704210381954908,
      "rewards/rejected": 0.00016899927868507802,
      "step": 160
    },
    {
      "epoch": 0.44,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": -2.0206995010375977,
      "logits/rejected": -2.0133931636810303,
      "logps/chosen": -33.504085540771484,
      "logps/rejected": -31.432220458984375,
      "loss": 0.9963,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.0029377774335443974,
      "rewards/margins": 0.003741443855687976,
      "rewards/rejected": -0.0008036663057282567,
      "step": 170
    },
    {
      "epoch": 0.47,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": -2.0269291400909424,
      "logits/rejected": -2.032160997390747,
      "logps/chosen": -32.24355697631836,
      "logps/rejected": -32.431182861328125,
      "loss": 0.9975,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.003012270200997591,
      "rewards/margins": 0.0024727012496441603,
      "rewards/rejected": 0.0005395688931457698,
      "step": 180
    },
    {
      "epoch": 0.49,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": -2.027367115020752,
      "logits/rejected": -2.0246078968048096,
      "logps/chosen": -31.290613174438477,
      "logps/rejected": -31.361133575439453,
      "loss": 0.9976,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.0019833946134895086,
      "rewards/margins": 0.002412599278613925,
      "rewards/rejected": -0.00042920451960526407,
      "step": 190
    },
    {
      "epoch": 0.52,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": -1.8985168933868408,
      "logits/rejected": -1.903148889541626,
      "logps/chosen": -31.30405616760254,
      "logps/rejected": -32.838443756103516,
      "loss": 0.9965,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.0030222723726183176,
      "rewards/margins": 0.0034973658621311188,
      "rewards/rejected": -0.0004750936641357839,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": -2.225177764892578,
      "eval_logits/rejected": -2.2203547954559326,
      "eval_logps/chosen": -34.06184387207031,
      "eval_logps/rejected": -37.579010009765625,
      "eval_loss": 0.9996482133865356,
      "eval_rewards/accuracies": 0.5070598125457764,
      "eval_rewards/chosen": -0.0002729461120907217,
      "eval_rewards/margins": 0.00035095339990220964,
      "eval_rewards/rejected": -0.0006238995119929314,
      "eval_runtime": 145.704,
      "eval_samples_per_second": 2.354,
      "eval_steps_per_second": 0.295,
      "step": 200
    },
    {
      "epoch": 0.55,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": -2.011120557785034,
      "logits/rejected": -2.021751880645752,
      "logps/chosen": -31.745685577392578,
      "logps/rejected": -33.96772003173828,
      "loss": 0.9967,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.0022561827208846807,
      "rewards/margins": 0.003306365106254816,
      "rewards/rejected": -0.0010501822689548135,
      "step": 210
    },
    {
      "epoch": 0.57,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": -1.903857946395874,
      "logits/rejected": -1.918621301651001,
      "logps/chosen": -29.797290802001953,
      "logps/rejected": -31.628814697265625,
      "loss": 0.9962,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.0031425058841705322,
      "rewards/margins": 0.0038488968275487423,
      "rewards/rejected": -0.0007063907687552273,
      "step": 220
    },
    {
      "epoch": 0.6,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": -1.9593979120254517,
      "logits/rejected": -1.9633464813232422,
      "logps/chosen": -33.067623138427734,
      "logps/rejected": -31.64206886291504,
      "loss": 0.9956,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.003498472273349762,
      "rewards/margins": 0.004426136147230864,
      "rewards/rejected": -0.0009276636992581189,
      "step": 230
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": -1.9572632312774658,
      "logits/rejected": -1.9354908466339111,
      "logps/chosen": -33.843727111816406,
      "logps/rejected": -35.1453742980957,
      "loss": 0.995,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.0031002266332507133,
      "rewards/margins": 0.0050460235215723515,
      "rewards/rejected": -0.0019457967719063163,
      "step": 240
    },
    {
      "epoch": 0.65,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": -1.9997854232788086,
      "logits/rejected": -1.9964803457260132,
      "logps/chosen": -32.75019454956055,
      "logps/rejected": -36.28661346435547,
      "loss": 0.9976,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.0017487213481217623,
      "rewards/margins": 0.0023655896075069904,
      "rewards/rejected": -0.0006168682011775672,
      "step": 250
    },
    {
      "epoch": 0.68,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": -1.8673791885375977,
      "logits/rejected": -1.8649587631225586,
      "logps/chosen": -34.018226623535156,
      "logps/rejected": -35.539276123046875,
      "loss": 0.998,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.001701725646853447,
      "rewards/margins": 0.002048287307843566,
      "rewards/rejected": -0.0003465614281594753,
      "step": 260
    },
    {
      "epoch": 0.7,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": -1.8522275686264038,
      "logits/rejected": -1.849872350692749,
      "logps/chosen": -34.16339874267578,
      "logps/rejected": -31.845317840576172,
      "loss": 0.9969,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.0023420110810548067,
      "rewards/margins": 0.0030819105450063944,
      "rewards/rejected": -0.0007398994639515877,
      "step": 270
    },
    {
      "epoch": 0.73,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": -1.9549518823623657,
      "logits/rejected": -1.94447922706604,
      "logps/chosen": -35.027687072753906,
      "logps/rejected": -31.895471572875977,
      "loss": 0.9962,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.003289591521024704,
      "rewards/margins": 0.003834384260699153,
      "rewards/rejected": -0.0005447928560897708,
      "step": 280
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": -2.0498766899108887,
      "logits/rejected": -2.034980297088623,
      "logps/chosen": -30.72440528869629,
      "logps/rejected": -32.658695220947266,
      "loss": 0.9979,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.0019971781875938177,
      "rewards/margins": 0.002065772656351328,
      "rewards/rejected": -6.859400309622288e-05,
      "step": 290
    },
    {
      "epoch": 0.78,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": -1.9201946258544922,
      "logits/rejected": -1.9177051782608032,
      "logps/chosen": -32.3183479309082,
      "logps/rejected": -30.95510482788086,
      "loss": 0.9925,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.006087628658860922,
      "rewards/margins": 0.007470599375665188,
      "rewards/rejected": -0.0013829706003889441,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": -2.2213804721832275,
      "eval_logits/rejected": -2.216555118560791,
      "eval_logps/chosen": -34.083614349365234,
      "eval_logps/rejected": -37.60634994506836,
      "eval_loss": 0.999591052532196,
      "eval_rewards/accuracies": 0.559385359287262,
      "eval_rewards/chosen": -0.0004906260874122381,
      "eval_rewards/margins": 0.00040669209556654096,
      "eval_rewards/rejected": -0.0008973181829787791,
      "eval_runtime": 145.8707,
      "eval_samples_per_second": 2.351,
      "eval_steps_per_second": 0.295,
      "step": 300
    },
    {
      "epoch": 0.81,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": -1.9060642719268799,
      "logits/rejected": -1.9028133153915405,
      "logps/chosen": -31.319162368774414,
      "logps/rejected": -33.85043716430664,
      "loss": 0.9961,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.002841859357431531,
      "rewards/margins": 0.003924719989299774,
      "rewards/rejected": -0.0010828599333763123,
      "step": 310
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": -1.9553836584091187,
      "logits/rejected": -1.9432109594345093,
      "logps/chosen": -34.27588653564453,
      "logps/rejected": -33.672359466552734,
      "loss": 0.9955,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.0032019000500440598,
      "rewards/margins": 0.004537059459835291,
      "rewards/rejected": -0.0013351596426218748,
      "step": 320
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": -1.9905990362167358,
      "logits/rejected": -1.9891618490219116,
      "logps/chosen": -33.116233825683594,
      "logps/rejected": -32.55724334716797,
      "loss": 0.9955,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.0036955769173800945,
      "rewards/margins": 0.004472161643207073,
      "rewards/rejected": -0.0007765850750729442,
      "step": 330
    },
    {
      "epoch": 0.88,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": -2.0769362449645996,
      "logits/rejected": -2.0613036155700684,
      "logps/chosen": -33.791297912597656,
      "logps/rejected": -33.12422180175781,
      "loss": 0.9961,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.003880967851728201,
      "rewards/margins": 0.00394281093031168,
      "rewards/rejected": -6.184288213262334e-05,
      "step": 340
    },
    {
      "epoch": 0.91,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": -1.950060248374939,
      "logits/rejected": -1.9492241144180298,
      "logps/chosen": -32.82404327392578,
      "logps/rejected": -32.50709915161133,
      "loss": 0.995,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.004580510314553976,
      "rewards/margins": 0.005000022705644369,
      "rewards/rejected": -0.000419511750806123,
      "step": 350
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": -1.9050449132919312,
      "logits/rejected": -1.915305733680725,
      "logps/chosen": -31.87860679626465,
      "logps/rejected": -35.34981155395508,
      "loss": 0.9961,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.0032608420588076115,
      "rewards/margins": 0.0038713677786290646,
      "rewards/rejected": -0.0006105261854827404,
      "step": 360
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": -2.04546856880188,
      "logits/rejected": -2.039043426513672,
      "logps/chosen": -33.336219787597656,
      "logps/rejected": -29.269311904907227,
      "loss": 0.9964,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.0031574335880577564,
      "rewards/margins": 0.003586276201531291,
      "rewards/rejected": -0.0004288425261620432,
      "step": 370
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": -1.905160665512085,
      "logits/rejected": -1.907360315322876,
      "logps/chosen": -33.86741256713867,
      "logps/rejected": -30.982807159423828,
      "loss": 0.9952,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.0037163912784308195,
      "rewards/margins": 0.004818186163902283,
      "rewards/rejected": -0.0011017953511327505,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.9973225085766284,
      "train_runtime": 3253.1307,
      "train_samples_per_second": 0.946,
      "train_steps_per_second": 0.118
    }
  ],
  "logging_steps": 10,
  "max_steps": 385,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}