{
  "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-08,
      "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": 1.2820512820512818e-07,
      "logits/chosen": -1.8661351203918457,
      "logits/rejected": -1.8704447746276855,
      "logps/chosen": -36.994537353515625,
      "logps/rejected": -33.65791320800781,
      "loss": 0.98,
      "rewards/accuracies": 0.5138888955116272,
      "rewards/chosen": 0.006032215431332588,
      "rewards/margins": 0.019952965900301933,
      "rewards/rejected": -0.013920750468969345,
      "step": 10
    },
    {
      "epoch": 0.05,
      "learning_rate": 2.5641025641025636e-07,
      "logits/chosen": -1.9972000122070312,
      "logits/rejected": -1.9998512268066406,
      "logps/chosen": -29.651025772094727,
      "logps/rejected": -29.047237396240234,
      "loss": 1.0125,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": -0.004414811730384827,
      "rewards/margins": -0.012516966089606285,
      "rewards/rejected": 0.008102154359221458,
      "step": 20
    },
    {
      "epoch": 0.08,
      "learning_rate": 3.8461538461538463e-07,
      "logits/chosen": -1.919605016708374,
      "logits/rejected": -1.9169189929962158,
      "logps/chosen": -31.410541534423828,
      "logps/rejected": -33.22766876220703,
      "loss": 0.9926,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.002776858163997531,
      "rewards/margins": 0.007382071111351252,
      "rewards/rejected": -0.004605212714523077,
      "step": 30
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.999896948438433e-07,
      "logits/chosen": -2.016796827316284,
      "logits/rejected": -2.0080485343933105,
      "logps/chosen": -32.59518814086914,
      "logps/rejected": -32.49793243408203,
      "loss": 1.0166,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.00919206440448761,
      "rewards/margins": -0.016562189906835556,
      "rewards/rejected": 0.007370126433670521,
      "step": 40
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.987541037542186e-07,
      "logits/chosen": -1.8643512725830078,
      "logits/rejected": -1.853578805923462,
      "logps/chosen": -33.55515670776367,
      "logps/rejected": -35.426795959472656,
      "loss": 1.0116,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": 0.0011615634430199862,
      "rewards/margins": -0.011613734066486359,
      "rewards/rejected": 0.012775297276675701,
      "step": 50
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.954691471941118e-07,
      "logits/chosen": -1.9452184438705444,
      "logits/rejected": -1.9471704959869385,
      "logps/chosen": -32.57328414916992,
      "logps/rejected": -33.17564010620117,
      "loss": 0.9844,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.013449115678668022,
      "rewards/margins": 0.015578977763652802,
      "rewards/rejected": -0.002129861619323492,
      "step": 60
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.901618883413548e-07,
      "logits/chosen": -2.079322338104248,
      "logits/rejected": -2.0843043327331543,
      "logps/chosen": -33.998104095458984,
      "logps/rejected": -36.586326599121094,
      "loss": 0.9961,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.004809260368347168,
      "rewards/margins": 0.003915563225746155,
      "rewards/rejected": -0.008724823594093323,
      "step": 70
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.828760511501322e-07,
      "logits/chosen": -1.9419012069702148,
      "logits/rejected": -1.9450546503067017,
      "logps/chosen": -34.40644836425781,
      "logps/rejected": -34.56013107299805,
      "loss": 1.0055,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.010679833590984344,
      "rewards/margins": -0.0054560392163693905,
      "rewards/rejected": 0.016135873273015022,
      "step": 80
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.736716601303429e-07,
      "logits/chosen": -1.9507659673690796,
      "logits/rejected": -1.9552879333496094,
      "logps/chosen": -32.45410919189453,
      "logps/rejected": -32.34144973754883,
      "loss": 1.0021,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.006657355930656195,
      "rewards/margins": -0.002072322415187955,
      "rewards/rejected": 0.008729678578674793,
      "step": 90
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.62624545834521e-07,
      "logits/chosen": -2.0492916107177734,
      "logits/rejected": -2.047295093536377,
      "logps/chosen": -32.23640441894531,
      "logps/rejected": -31.247751235961914,
      "loss": 1.007,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": 0.002216553781181574,
      "rewards/margins": -0.007013450376689434,
      "rewards/rejected": 0.009230004623532295,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": -2.2440249919891357,
      "eval_logits/rejected": -2.239149808883667,
      "eval_logps/chosen": -34.017887115478516,
      "eval_logps/rejected": -37.50767517089844,
      "eval_loss": 0.9963483810424805,
      "eval_rewards/accuracies": 0.5166113376617432,
      "eval_rewards/chosen": 0.008332207798957825,
      "eval_rewards/margins": 0.0038595683872699738,
      "eval_rewards/rejected": 0.004472639411687851,
      "eval_runtime": 146.1438,
      "eval_samples_per_second": 2.347,
      "eval_steps_per_second": 0.294,
      "step": 100
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.4982572012636904e-07,
      "logits/chosen": -2.005338191986084,
      "logits/rejected": -2.002920627593994,
      "logps/chosen": -33.22484588623047,
      "logps/rejected": -34.01424026489258,
      "loss": 0.9981,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": 0.009485405869781971,
      "rewards/margins": 0.001919901347719133,
      "rewards/rejected": 0.007565504405647516,
      "step": 110
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.353806263777677e-07,
      "logits/chosen": -2.0164568424224854,
      "logits/rejected": -2.008070468902588,
      "logps/chosen": -32.454193115234375,
      "logps/rejected": -32.175106048583984,
      "loss": 0.9978,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": -0.004645171575248241,
      "rewards/margins": 0.0021557719446718693,
      "rewards/rejected": -0.006800943519920111,
      "step": 120
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.194082707715275e-07,
      "logits/chosen": -2.046295642852783,
      "logits/rejected": -2.03825044631958,
      "logps/chosen": -30.491985321044922,
      "logps/rejected": -32.04901885986328,
      "loss": 1.0101,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": -0.008134904317557812,
      "rewards/margins": -0.010127933695912361,
      "rewards/rejected": 0.001993028912693262,
      "step": 130
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.020402418666621e-07,
      "logits/chosen": -1.9772052764892578,
      "logits/rejected": -1.9874699115753174,
      "logps/chosen": -31.385913848876953,
      "logps/rejected": -32.56010818481445,
      "loss": 0.9705,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.016873350366950035,
      "rewards/margins": 0.02948569692671299,
      "rewards/rejected": -0.01261234562844038,
      "step": 140
    },
    {
      "epoch": 0.39,
      "learning_rate": 3.8341962650351185e-07,
      "logits/chosen": -1.8907018899917603,
      "logits/rejected": -1.8917919397354126,
      "logps/chosen": -34.191062927246094,
      "logps/rejected": -34.757171630859375,
      "loss": 0.9992,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.0008138298871926963,
      "rewards/margins": 0.0007720142602920532,
      "rewards/rejected": 4.181563781457953e-05,
      "step": 150
    },
    {
      "epoch": 0.42,
      "learning_rate": 3.636998309800572e-07,
      "logits/chosen": -1.9424282312393188,
      "logits/rejected": -1.9389431476593018,
      "logps/chosen": -36.157745361328125,
      "logps/rejected": -32.73271942138672,
      "loss": 0.988,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.0069676609709858894,
      "rewards/margins": 0.011951583437621593,
      "rewards/rejected": -0.004983922932296991,
      "step": 160
    },
    {
      "epoch": 0.44,
      "learning_rate": 3.430433172111807e-07,
      "logits/chosen": -2.0421340465545654,
      "logits/rejected": -2.0347371101379395,
      "logps/chosen": -33.798614501953125,
      "logps/rejected": -31.370548248291016,
      "loss": 0.991,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.00037657321081496775,
      "rewards/margins": 0.008971447125077248,
      "rewards/rejected": -0.00934801995754242,
      "step": 170
    },
    {
      "epoch": 0.47,
      "learning_rate": 3.216202642830543e-07,
      "logits/chosen": -2.0471675395965576,
      "logits/rejected": -2.0524353981018066,
      "logps/chosen": -32.519779205322266,
      "logps/rejected": -32.49123001098633,
      "loss": 0.9845,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": 0.012504413723945618,
      "rewards/margins": 0.015548673458397388,
      "rewards/rejected": -0.003044259501621127,
      "step": 180
    },
    {
      "epoch": 0.49,
      "learning_rate": 2.9960716642946403e-07,
      "logits/chosen": -2.0482120513916016,
      "logits/rejected": -2.0454375743865967,
      "logps/chosen": -31.485937118530273,
      "logps/rejected": -31.332433700561523,
      "loss": 0.9914,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.0015064775943756104,
      "rewards/margins": 0.008616042323410511,
      "rewards/rejected": -0.007109564729034901,
      "step": 190
    },
    {
      "epoch": 0.52,
      "learning_rate": 2.771853789806683e-07,
      "logits/chosen": -1.9187748432159424,
      "logits/rejected": -1.9234533309936523,
      "logps/chosen": -31.564071655273438,
      "logps/rejected": -32.80377960205078,
      "loss": 0.9725,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.021104933694005013,
      "rewards/margins": 0.027528975158929825,
      "rewards/rejected": -0.006424039602279663,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": -2.24338698387146,
      "eval_logits/rejected": -2.2385094165802,
      "eval_logps/chosen": -34.032955169677734,
      "eval_logps/rejected": -37.49824142456055,
      "eval_loss": 1.0085184574127197,
      "eval_rewards/accuracies": 0.4991694688796997,
      "eval_rewards/chosen": 0.0007996229687705636,
      "eval_rewards/margins": -0.008389986120164394,
      "eval_rewards/rejected": 0.009189609438180923,
      "eval_runtime": 145.9067,
      "eval_samples_per_second": 2.351,
      "eval_steps_per_second": 0.295,
      "step": 200
    },
    {
      "epoch": 0.55,
      "learning_rate": 2.5453962426402e-07,
      "logits/chosen": -2.0320358276367188,
      "logits/rejected": -2.042719602584839,
      "logps/chosen": -31.940053939819336,
      "logps/rejected": -33.88956069946289,
      "loss": 0.9709,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.015625199303030968,
      "rewards/margins": 0.029055744409561157,
      "rewards/rejected": -0.01343054324388504,
      "step": 210
    },
    {
      "epoch": 0.57,
      "learning_rate": 2.318564697655179e-07,
      "logits/chosen": -1.9255237579345703,
      "logits/rejected": -1.9403839111328125,
      "logps/chosen": -30.091089248657227,
      "logps/rejected": -31.54927635192871,
      "loss": 0.9942,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.010226013138890266,
      "rewards/margins": 0.005776937119662762,
      "rewards/rejected": 0.004449075553566217,
      "step": 220
    },
    {
      "epoch": 0.6,
      "learning_rate": 2.093227910899832e-07,
      "logits/chosen": -1.9833370447158813,
      "logits/rejected": -1.9873158931732178,
      "logps/chosen": -33.393131256103516,
      "logps/rejected": -31.559478759765625,
      "loss": 0.9827,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": 0.012168514542281628,
      "rewards/margins": 0.017256852239370346,
      "rewards/rejected": -0.005088338162750006,
      "step": 230
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.8712423238279356e-07,
      "logits/chosen": -1.9829031229019165,
      "logits/rejected": -1.9609248638153076,
      "logps/chosen": -34.161617279052734,
      "logps/rejected": -34.973785400390625,
      "loss": 0.9924,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.003932595252990723,
      "rewards/margins": 0.007563352584838867,
      "rewards/rejected": -0.01149594783782959,
      "step": 240
    },
    {
      "epoch": 0.65,
      "learning_rate": 1.654436768970182e-07,
      "logits/chosen": -2.0244410037994385,
      "logits/rejected": -2.0211281776428223,
      "logps/chosen": -32.90896987915039,
      "logps/rejected": -36.21443557739258,
      "loss": 0.9972,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": 0.008049068041145802,
      "rewards/margins": 0.0028051852714270353,
      "rewards/rejected": 0.0052438825368881226,
      "step": 250
    },
    {
      "epoch": 0.68,
      "learning_rate": 1.444597403062196e-07,
      "logits/chosen": -1.8918297290802002,
      "logits/rejected": -1.8893934488296509,
      "logps/chosen": -34.178001403808594,
      "logps/rejected": -35.52972412109375,
      "loss": 0.9823,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.00519949197769165,
      "rewards/margins": 0.017749812453985214,
      "rewards/rejected": -0.012550321407616138,
      "step": 260
    },
    {
      "epoch": 0.7,
      "learning_rate": 1.2434529917578887e-07,
      "logits/chosen": -1.875998854637146,
      "logits/rejected": -1.873445749282837,
      "logps/chosen": -34.39997100830078,
      "logps/rejected": -31.749908447265625,
      "loss": 1.0119,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": -0.0011831462616100907,
      "rewards/margins": -0.011894017457962036,
      "rewards/rejected": 0.01071087084710598,
      "step": 270
    },
    {
      "epoch": 0.73,
      "learning_rate": 1.0526606671603521e-07,
      "logits/chosen": -1.9797567129135132,
      "logits/rejected": -1.9691368341445923,
      "logps/chosen": -35.324066162109375,
      "logps/rejected": -31.857070922851562,
      "loss": 0.9757,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.016287770122289658,
      "rewards/margins": 0.024326175451278687,
      "rewards/rejected": -0.008038404397666454,
      "step": 280
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.737922755071453e-08,
      "logits/chosen": -2.0752615928649902,
      "logits/rejected": -2.06022310256958,
      "logps/chosen": -30.904525756835938,
      "logps/rejected": -32.66613006591797,
      "loss": 0.9831,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.00980077963322401,
      "rewards/margins": 0.016948198899626732,
      "rewards/rejected": -0.007147419266402721,
      "step": 290
    },
    {
      "epoch": 0.78,
      "learning_rate": 7.08321427484816e-08,
      "logits/chosen": -1.946509599685669,
      "logits/rejected": -1.9439618587493896,
      "logps/chosen": -32.89145278930664,
      "logps/rejected": -30.81719398498535,
      "loss": 0.982,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.017829114571213722,
      "rewards/margins": 0.018021035939455032,
      "rewards/rejected": -0.00019192098989151418,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": -2.2438137531280518,
      "eval_logits/rejected": -2.2389373779296875,
      "eval_logps/chosen": -34.02228546142578,
      "eval_logps/rejected": -37.50155258178711,
      "eval_loss": 1.0014283657073975,
      "eval_rewards/accuracies": 0.5103820562362671,
      "eval_rewards/chosen": 0.006133119110018015,
      "eval_rewards/margins": -0.001401092391461134,
      "eval_rewards/rejected": 0.007534211501479149,
      "eval_runtime": 145.9391,
      "eval_samples_per_second": 2.35,
      "eval_steps_per_second": 0.295,
      "step": 300
    },
    {
      "epoch": 0.81,
      "learning_rate": 5.576113578589034e-08,
      "logits/chosen": -1.9287340641021729,
      "logits/rejected": -1.9254703521728516,
      "logps/chosen": -31.576568603515625,
      "logps/rejected": -33.73060607910156,
      "loss": 0.9924,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.013391390442848206,
      "rewards/margins": 0.00761887151747942,
      "rewards/rejected": 0.005772518925368786,
      "step": 310
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.229036944380912e-08,
      "logits/chosen": -1.980047583580017,
      "logits/rejected": -1.9677263498306274,
      "logps/chosen": -34.565635681152344,
      "logps/rejected": -33.560508728027344,
      "loss": 0.9739,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.015218675136566162,
      "rewards/margins": 0.02605234459042549,
      "rewards/rejected": -0.010833668522536755,
      "step": 320
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.053082288996112e-08,
      "logits/chosen": -2.0154004096984863,
      "logits/rejected": -2.013948917388916,
      "logps/chosen": -33.46211624145508,
      "logps/rejected": -32.472103118896484,
      "loss": 0.9919,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.011838942766189575,
      "rewards/margins": 0.008100616745650768,
      "rewards/rejected": 0.0037383257877081633,
      "step": 330
    },
    {
      "epoch": 0.88,
      "learning_rate": 2.05793773749158e-08,
      "logits/chosen": -2.1025614738464355,
      "logits/rejected": -2.086764097213745,
      "logps/chosen": -34.168495178222656,
      "logps/rejected": -33.07737731933594,
      "loss": 1.0149,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": 0.005447447299957275,
      "rewards/margins": -0.014881876297295094,
      "rewards/rejected": 0.020329322665929794,
      "step": 340
    },
    {
      "epoch": 0.91,
      "learning_rate": 1.251801807404168e-08,
      "logits/chosen": -1.9740177392959595,
      "logits/rejected": -1.9730663299560547,
      "logps/chosen": -33.242271423339844,
      "logps/rejected": -32.460182189941406,
      "loss": 0.9826,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.019913053140044212,
      "rewards/margins": 0.017432674765586853,
      "rewards/rejected": 0.0024803788401186466,
      "step": 350
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.41315865106129e-09,
      "logits/chosen": -1.9304364919662476,
      "logits/rejected": -1.9407978057861328,
      "logps/chosen": -32.19029998779297,
      "logps/rejected": -35.303955078125,
      "loss": 0.9852,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.007193857338279486,
      "rewards/margins": 0.014792281202971935,
      "rewards/rejected": -0.00759842386469245,
      "step": 360
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.3150941078050324e-09,
      "logits/chosen": -2.069343090057373,
      "logits/rejected": -2.062788724899292,
      "logps/chosen": -33.646358489990234,
      "logps/rejected": -29.213809967041016,
      "loss": 1.0035,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": 0.0028047324158251286,
      "rewards/margins": -0.0035038769710808992,
      "rewards/rejected": 0.006308609154075384,
      "step": 370
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.575864278703266e-10,
      "logits/chosen": -1.9290740489959717,
      "logits/rejected": -1.9312365055084229,
      "logps/chosen": -34.245155334472656,
      "logps/rejected": -30.889087677001953,
      "loss": 0.9948,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": -0.003053072141483426,
      "rewards/margins": 0.005176535341888666,
      "rewards/rejected": -0.008229607716202736,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.9930156608680626,
      "train_runtime": 3255.7181,
      "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
}