{
  "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,
      "grad_norm": 13.0625,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": 88.18099975585938,
      "logits/rejected": 88.25153350830078,
      "logps/chosen": -29.073104858398438,
      "logps/rejected": -26.25731658935547,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 10.6875,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": 81.09551239013672,
      "logits/rejected": 80.80068969726562,
      "logps/chosen": -34.28557586669922,
      "logps/rejected": -33.081260681152344,
      "loss": 0.6879,
      "rewards/accuracies": 0.5277777910232544,
      "rewards/chosen": -0.008770154789090157,
      "rewards/margins": 0.01580061763525009,
      "rewards/rejected": -0.0245707705616951,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 11.8125,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": 80.68476867675781,
      "logits/rejected": 80.57356262207031,
      "logps/chosen": -33.56308364868164,
      "logps/rejected": -30.81569480895996,
      "loss": 0.6814,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.02536492981016636,
      "rewards/margins": 0.03300638496875763,
      "rewards/rejected": -0.007641455624252558,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 12.4375,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": 82.52334594726562,
      "logits/rejected": 82.55731964111328,
      "logps/chosen": -33.843528747558594,
      "logps/rejected": -31.111019134521484,
      "loss": 0.7142,
      "rewards/accuracies": 0.4124999940395355,
      "rewards/chosen": 0.03385842964053154,
      "rewards/margins": -0.028703877702355385,
      "rewards/rejected": 0.06256230920553207,
      "step": 30
    },
    {
      "epoch": 0.1,
      "grad_norm": 11.0,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": 81.0788803100586,
      "logits/rejected": 81.07527923583984,
      "logps/chosen": -32.68939971923828,
      "logps/rejected": -33.106483459472656,
      "loss": 0.6804,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.0969332829117775,
      "rewards/margins": 0.044524990022182465,
      "rewards/rejected": 0.05240829661488533,
      "step": 40
    },
    {
      "epoch": 0.13,
      "grad_norm": 9.25,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": 78.8182373046875,
      "logits/rejected": 78.82775115966797,
      "logps/chosen": -30.46297836303711,
      "logps/rejected": -30.7069091796875,
      "loss": 0.6746,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.14845404028892517,
      "rewards/margins": 0.07517841458320618,
      "rewards/rejected": 0.073275625705719,
      "step": 50
    },
    {
      "epoch": 0.16,
      "grad_norm": 10.0,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": 83.39994812011719,
      "logits/rejected": 83.45416259765625,
      "logps/chosen": -30.7622013092041,
      "logps/rejected": -29.376235961914062,
      "loss": 0.6858,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.08259173482656479,
      "rewards/margins": 0.03973449766635895,
      "rewards/rejected": 0.04285724088549614,
      "step": 60
    },
    {
      "epoch": 0.18,
      "grad_norm": 14.3125,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": 84.03162384033203,
      "logits/rejected": 84.06212615966797,
      "logps/chosen": -30.389270782470703,
      "logps/rejected": -32.93138885498047,
      "loss": 0.6894,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": 0.04795076698064804,
      "rewards/margins": 0.02668655477464199,
      "rewards/rejected": 0.02126421593129635,
      "step": 70
    },
    {
      "epoch": 0.21,
      "grad_norm": 12.0,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": 81.65167236328125,
      "logits/rejected": 81.63506317138672,
      "logps/chosen": -31.229572296142578,
      "logps/rejected": -30.818950653076172,
      "loss": 0.6715,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.05507441610097885,
      "rewards/margins": 0.06733567267656326,
      "rewards/rejected": -0.012261250987648964,
      "step": 80
    },
    {
      "epoch": 0.23,
      "grad_norm": 13.875,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": 78.45601654052734,
      "logits/rejected": 78.42174530029297,
      "logps/chosen": -32.425289154052734,
      "logps/rejected": -31.132177352905273,
      "loss": 0.6671,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.04257836565375328,
      "rewards/margins": 0.07871954143047333,
      "rewards/rejected": -0.03614116832613945,
      "step": 90
    },
    {
      "epoch": 0.26,
      "grad_norm": 11.4375,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": 83.59445190429688,
      "logits/rejected": 83.61743927001953,
      "logps/chosen": -33.955238342285156,
      "logps/rejected": -31.89117431640625,
      "loss": 0.6551,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.06995107233524323,
      "rewards/margins": 0.09807688742876053,
      "rewards/rejected": -0.028125811368227005,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": 98.74779510498047,
      "eval_logits/rejected": 98.73747253417969,
      "eval_logps/chosen": -32.46078872680664,
      "eval_logps/rejected": -36.02348327636719,
      "eval_loss": 0.6956869959831238,
      "eval_rewards/accuracies": 0.5224252939224243,
      "eval_rewards/chosen": -0.003522348590195179,
      "eval_rewards/margins": 0.00786901917308569,
      "eval_rewards/rejected": -0.011391367763280869,
      "eval_runtime": 104.3728,
      "eval_samples_per_second": 3.286,
      "eval_steps_per_second": 0.412,
      "step": 100
    },
    {
      "epoch": 0.29,
      "grad_norm": 14.0,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": 83.79670715332031,
      "logits/rejected": 83.66975402832031,
      "logps/chosen": -32.42261505126953,
      "logps/rejected": -32.766204833984375,
      "loss": 0.6376,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.12003795802593231,
      "rewards/margins": 0.1421007364988327,
      "rewards/rejected": -0.02206278033554554,
      "step": 110
    },
    {
      "epoch": 0.31,
      "grad_norm": 13.4375,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": 83.85148620605469,
      "logits/rejected": 83.95706939697266,
      "logps/chosen": -28.113927841186523,
      "logps/rejected": -35.665374755859375,
      "loss": 0.6109,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.1631302535533905,
      "rewards/margins": 0.20437581837177277,
      "rewards/rejected": -0.04124556854367256,
      "step": 120
    },
    {
      "epoch": 0.34,
      "grad_norm": 9.5,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": 80.99261474609375,
      "logits/rejected": 81.01200866699219,
      "logps/chosen": -30.389972686767578,
      "logps/rejected": -32.014381408691406,
      "loss": 0.6511,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.10347676277160645,
      "rewards/margins": 0.120028056204319,
      "rewards/rejected": -0.0165513064712286,
      "step": 130
    },
    {
      "epoch": 0.36,
      "grad_norm": 9.3125,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": 82.00664520263672,
      "logits/rejected": 82.01942443847656,
      "logps/chosen": -27.06869125366211,
      "logps/rejected": -33.10210418701172,
      "loss": 0.611,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.07572650164365768,
      "rewards/margins": 0.21147167682647705,
      "rewards/rejected": -0.13574519753456116,
      "step": 140
    },
    {
      "epoch": 0.39,
      "grad_norm": 10.0625,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": 80.3797378540039,
      "logits/rejected": 80.35098266601562,
      "logps/chosen": -28.888397216796875,
      "logps/rejected": -33.30742645263672,
      "loss": 0.596,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.10014281421899796,
      "rewards/margins": 0.24832835793495178,
      "rewards/rejected": -0.14818556606769562,
      "step": 150
    },
    {
      "epoch": 0.42,
      "grad_norm": 13.9375,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": 82.10334014892578,
      "logits/rejected": 82.10926818847656,
      "logps/chosen": -33.88823699951172,
      "logps/rejected": -30.74798011779785,
      "loss": 0.6209,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.04970277473330498,
      "rewards/margins": 0.2087908238172531,
      "rewards/rejected": -0.15908804535865784,
      "step": 160
    },
    {
      "epoch": 0.44,
      "grad_norm": 12.4375,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": 82.82877349853516,
      "logits/rejected": 82.77672576904297,
      "logps/chosen": -30.899311065673828,
      "logps/rejected": -32.958866119384766,
      "loss": 0.598,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.0719953253865242,
      "rewards/margins": 0.2597564160823822,
      "rewards/rejected": -0.1877611130475998,
      "step": 170
    },
    {
      "epoch": 0.47,
      "grad_norm": 8.75,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": 80.19332885742188,
      "logits/rejected": 80.17335510253906,
      "logps/chosen": -30.60544204711914,
      "logps/rejected": -31.953319549560547,
      "loss": 0.5849,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.1183418408036232,
      "rewards/margins": 0.2775646448135376,
      "rewards/rejected": -0.159222811460495,
      "step": 180
    },
    {
      "epoch": 0.49,
      "grad_norm": 6.59375,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": 81.75331115722656,
      "logits/rejected": 81.74398040771484,
      "logps/chosen": -30.514236450195312,
      "logps/rejected": -30.920120239257812,
      "loss": 0.6564,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.050435226410627365,
      "rewards/margins": 0.14085446298122406,
      "rewards/rejected": -0.090419240295887,
      "step": 190
    },
    {
      "epoch": 0.52,
      "grad_norm": 8.9375,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": 77.08003997802734,
      "logits/rejected": 77.01986694335938,
      "logps/chosen": -33.93625259399414,
      "logps/rejected": -33.12162780761719,
      "loss": 0.5948,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.16329774260520935,
      "rewards/margins": 0.28894227743148804,
      "rewards/rejected": -0.1256445348262787,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": 98.33805847167969,
      "eval_logits/rejected": 98.30875396728516,
      "eval_logps/chosen": -32.75787353515625,
      "eval_logps/rejected": -36.611534118652344,
      "eval_loss": 0.672012209892273,
      "eval_rewards/accuracies": 0.5514950156211853,
      "eval_rewards/chosen": -0.06293957680463791,
      "eval_rewards/margins": 0.06606128066778183,
      "eval_rewards/rejected": -0.12900085747241974,
      "eval_runtime": 104.1536,
      "eval_samples_per_second": 3.293,
      "eval_steps_per_second": 0.413,
      "step": 200
    },
    {
      "epoch": 0.55,
      "grad_norm": 15.0625,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": 79.71280670166016,
      "logits/rejected": 79.62548065185547,
      "logps/chosen": -33.420021057128906,
      "logps/rejected": -35.76679229736328,
      "loss": 0.6026,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.10514093935489655,
      "rewards/margins": 0.2612799406051636,
      "rewards/rejected": -0.15613897144794464,
      "step": 210
    },
    {
      "epoch": 0.57,
      "grad_norm": 10.375,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": 81.73656463623047,
      "logits/rejected": 81.82989501953125,
      "logps/chosen": -31.172740936279297,
      "logps/rejected": -31.53964614868164,
      "loss": 0.5726,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": 0.15524254739284515,
      "rewards/margins": 0.3173803985118866,
      "rewards/rejected": -0.16213785111904144,
      "step": 220
    },
    {
      "epoch": 0.6,
      "grad_norm": 11.5625,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": 78.81214904785156,
      "logits/rejected": 78.8668212890625,
      "logps/chosen": -32.4487190246582,
      "logps/rejected": -34.767303466796875,
      "loss": 0.6172,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.07140493392944336,
      "rewards/margins": 0.22243526577949524,
      "rewards/rejected": -0.15103033185005188,
      "step": 230
    },
    {
      "epoch": 0.62,
      "grad_norm": 14.25,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": 81.07857513427734,
      "logits/rejected": 81.38665771484375,
      "logps/chosen": -31.030254364013672,
      "logps/rejected": -32.145328521728516,
      "loss": 0.5977,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.13156220316886902,
      "rewards/margins": 0.254219114780426,
      "rewards/rejected": -0.12265688180923462,
      "step": 240
    },
    {
      "epoch": 0.65,
      "grad_norm": 12.75,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": 79.71233367919922,
      "logits/rejected": 79.77317810058594,
      "logps/chosen": -27.3262939453125,
      "logps/rejected": -30.594249725341797,
      "loss": 0.6333,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.053920842707157135,
      "rewards/margins": 0.19262339174747467,
      "rewards/rejected": -0.13870255649089813,
      "step": 250
    },
    {
      "epoch": 0.68,
      "grad_norm": 10.5,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": 76.8447036743164,
      "logits/rejected": 76.98011779785156,
      "logps/chosen": -30.586124420166016,
      "logps/rejected": -37.12653350830078,
      "loss": 0.5519,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.19497378170490265,
      "rewards/margins": 0.3999541699886322,
      "rewards/rejected": -0.20498037338256836,
      "step": 260
    },
    {
      "epoch": 0.7,
      "grad_norm": 9.3125,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": 76.1262435913086,
      "logits/rejected": 76.15739440917969,
      "logps/chosen": -31.1810245513916,
      "logps/rejected": -32.34056854248047,
      "loss": 0.5874,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.11957981437444687,
      "rewards/margins": 0.28660956025123596,
      "rewards/rejected": -0.1670297235250473,
      "step": 270
    },
    {
      "epoch": 0.73,
      "grad_norm": 15.75,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": 78.95848846435547,
      "logits/rejected": 78.7259750366211,
      "logps/chosen": -31.400470733642578,
      "logps/rejected": -30.376489639282227,
      "loss": 0.6251,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.08188416808843613,
      "rewards/margins": 0.22378277778625488,
      "rewards/rejected": -0.14189860224723816,
      "step": 280
    },
    {
      "epoch": 0.75,
      "grad_norm": 9.3125,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": 79.04127502441406,
      "logits/rejected": 78.95845031738281,
      "logps/chosen": -33.34519577026367,
      "logps/rejected": -33.35126876831055,
      "loss": 0.5388,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": 0.15181000530719757,
      "rewards/margins": 0.4186907410621643,
      "rewards/rejected": -0.26688069105148315,
      "step": 290
    },
    {
      "epoch": 0.78,
      "grad_norm": 12.8125,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": 74.71063995361328,
      "logits/rejected": 74.8211441040039,
      "logps/chosen": -32.499046325683594,
      "logps/rejected": -29.784500122070312,
      "loss": 0.5903,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.18133236467838287,
      "rewards/margins": 0.3178277611732483,
      "rewards/rejected": -0.13649536669254303,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": 98.2787857055664,
      "eval_logits/rejected": 98.24710083007812,
      "eval_logps/chosen": -32.80708312988281,
      "eval_logps/rejected": -36.62175750732422,
      "eval_loss": 0.6774110198020935,
      "eval_rewards/accuracies": 0.5340532064437866,
      "eval_rewards/chosen": -0.07278140634298325,
      "eval_rewards/margins": 0.05826449766755104,
      "eval_rewards/rejected": -0.13104592263698578,
      "eval_runtime": 104.1558,
      "eval_samples_per_second": 3.293,
      "eval_steps_per_second": 0.413,
      "step": 300
    },
    {
      "epoch": 0.81,
      "grad_norm": 10.0,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": 81.97877502441406,
      "logits/rejected": 82.00166320800781,
      "logps/chosen": -30.284442901611328,
      "logps/rejected": -33.162872314453125,
      "loss": 0.579,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.11581814289093018,
      "rewards/margins": 0.30844008922576904,
      "rewards/rejected": -0.19262196123600006,
      "step": 310
    },
    {
      "epoch": 0.83,
      "grad_norm": 9.8125,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": 79.28789520263672,
      "logits/rejected": 79.29541778564453,
      "logps/chosen": -30.684707641601562,
      "logps/rejected": -29.662694931030273,
      "loss": 0.5734,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.1827705204486847,
      "rewards/margins": 0.3314109444618225,
      "rewards/rejected": -0.14864039421081543,
      "step": 320
    },
    {
      "epoch": 0.86,
      "grad_norm": 9.4375,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": 76.40345764160156,
      "logits/rejected": 76.45513916015625,
      "logps/chosen": -29.180185317993164,
      "logps/rejected": -33.6518440246582,
      "loss": 0.5234,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.2344847470521927,
      "rewards/margins": 0.43788036704063416,
      "rewards/rejected": -0.20339563488960266,
      "step": 330
    },
    {
      "epoch": 0.88,
      "grad_norm": 16.875,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": 80.8781509399414,
      "logits/rejected": 80.92503356933594,
      "logps/chosen": -32.53740310668945,
      "logps/rejected": -34.52736282348633,
      "loss": 0.5705,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.13494810461997986,
      "rewards/margins": 0.3666823208332062,
      "rewards/rejected": -0.23173420131206512,
      "step": 340
    },
    {
      "epoch": 0.91,
      "grad_norm": 9.375,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": 79.81520080566406,
      "logits/rejected": 79.83287048339844,
      "logps/chosen": -32.53202819824219,
      "logps/rejected": -33.971153259277344,
      "loss": 0.561,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.22234222292900085,
      "rewards/margins": 0.39821282029151917,
      "rewards/rejected": -0.1758706122636795,
      "step": 350
    },
    {
      "epoch": 0.94,
      "grad_norm": 9.25,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": 81.41835021972656,
      "logits/rejected": 81.45225524902344,
      "logps/chosen": -28.548419952392578,
      "logps/rejected": -32.2237434387207,
      "loss": 0.5819,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.20075836777687073,
      "rewards/margins": 0.3155379891395569,
      "rewards/rejected": -0.11477959156036377,
      "step": 360
    },
    {
      "epoch": 0.96,
      "grad_norm": 12.4375,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": 80.87324523925781,
      "logits/rejected": 80.88945007324219,
      "logps/chosen": -32.29994583129883,
      "logps/rejected": -36.0130615234375,
      "loss": 0.6003,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.1152433529496193,
      "rewards/margins": 0.295980840921402,
      "rewards/rejected": -0.1807374656200409,
      "step": 370
    },
    {
      "epoch": 0.99,
      "grad_norm": 11.1875,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": 74.64203643798828,
      "logits/rejected": 74.50636291503906,
      "logps/chosen": -29.965755462646484,
      "logps/rejected": -28.94148826599121,
      "loss": 0.5896,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.12207402288913727,
      "rewards/margins": 0.2782566249370575,
      "rewards/rejected": -0.15618260204792023,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.6168378699909557,
      "train_runtime": 2560.1818,
      "train_samples_per_second": 1.203,
      "train_steps_per_second": 0.15
    }
  ],
  "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
}