{
  "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": 0.3086,
      "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": -1.866453766822815,
      "logits/rejected": -1.8707627058029175,
      "logps/chosen": -36.98646545410156,
      "logps/rejected": -33.6445426940918,
      "loss": 0.3098,
      "rewards/accuracies": 0.4444444477558136,
      "rewards/chosen": 0.018122652545571327,
      "rewards/margins": 0.031144678592681885,
      "rewards/rejected": -0.013022023253142834,
      "step": 10
    },
    {
      "epoch": 0.05,
      "learning_rate": 2.5641025641025636e-07,
      "logits/chosen": -1.9970738887786865,
      "logits/rejected": -1.9997297525405884,
      "logps/chosen": -29.651851654052734,
      "logps/rejected": -29.048221588134766,
      "loss": 0.3674,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.008690872229635715,
      "rewards/margins": -0.02239001728594303,
      "rewards/rejected": 0.01369914598762989,
      "step": 20
    },
    {
      "epoch": 0.08,
      "learning_rate": 3.8461538461538463e-07,
      "logits/chosen": -1.919904351234436,
      "logits/rejected": -1.9172093868255615,
      "logps/chosen": -31.407363891601562,
      "logps/rejected": -33.22914505004883,
      "loss": 0.3289,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.007855755276978016,
      "rewards/margins": 0.017470674589276314,
      "rewards/rejected": -0.009614917449653149,
      "step": 30
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.999896948438433e-07,
      "logits/chosen": -2.016488552093506,
      "logits/rejected": -2.0077414512634277,
      "logps/chosen": -32.588050842285156,
      "logps/rejected": -32.51618576049805,
      "loss": 0.3526,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": -0.010120267979800701,
      "rewards/margins": -0.006958496756851673,
      "rewards/rejected": -0.0031617730855941772,
      "step": 40
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.987541037542186e-07,
      "logits/chosen": -1.8645082712173462,
      "logits/rejected": -1.8537276983261108,
      "logps/chosen": -33.5635871887207,
      "logps/rejected": -35.42169952392578,
      "loss": 0.3858,
      "rewards/accuracies": 0.38749998807907104,
      "rewards/chosen": -0.005497059319168329,
      "rewards/margins": -0.03307711333036423,
      "rewards/rejected": 0.027580058202147484,
      "step": 50
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.954691471941118e-07,
      "logits/chosen": -1.9458835124969482,
      "logits/rejected": -1.9478065967559814,
      "logps/chosen": -32.57460021972656,
      "logps/rejected": -33.195518493652344,
      "loss": 0.3015,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.02302517369389534,
      "rewards/margins": 0.04474987834692001,
      "rewards/rejected": -0.021724706515669823,
      "step": 60
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.901618883413548e-07,
      "logits/chosen": -2.0793328285217285,
      "logits/rejected": -2.084324836730957,
      "logps/chosen": -33.99799728393555,
      "logps/rejected": -36.57799530029297,
      "loss": 0.3658,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": -0.00855917576700449,
      "rewards/margins": -0.00035296304849907756,
      "rewards/rejected": -0.008206211030483246,
      "step": 70
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.828760511501322e-07,
      "logits/chosen": -1.9421542882919312,
      "logits/rejected": -1.9453080892562866,
      "logps/chosen": -34.38970184326172,
      "logps/rejected": -34.58138656616211,
      "loss": 0.3181,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.034300822764635086,
      "rewards/margins": 0.024381538853049278,
      "rewards/rejected": 0.009919276461005211,
      "step": 80
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.736716601303429e-07,
      "logits/chosen": -1.9507806301116943,
      "logits/rejected": -1.955293893814087,
      "logps/chosen": -32.461936950683594,
      "logps/rejected": -32.3816032409668,
      "loss": 0.3094,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.004934861324727535,
      "rewards/margins": 0.02535998821258545,
      "rewards/rejected": -0.02042512409389019,
      "step": 90
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.62624545834521e-07,
      "logits/chosen": -2.0485286712646484,
      "logits/rejected": -2.046539783477783,
      "logps/chosen": -32.23391342163086,
      "logps/rejected": -31.265544891357422,
      "loss": 0.333,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": 0.006228716112673283,
      "rewards/margins": 0.005629505030810833,
      "rewards/rejected": 0.0005992107326164842,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": -2.243820905685425,
      "eval_logits/rejected": -2.238938570022583,
      "eval_logps/chosen": -34.02050018310547,
      "eval_logps/rejected": -37.50347137451172,
      "eval_loss": 0.3404548168182373,
      "eval_rewards/accuracies": 0.5103820562362671,
      "eval_rewards/chosen": 0.012649022974073887,
      "eval_rewards/margins": 0.0008147989865392447,
      "eval_rewards/rejected": 0.011834223754703999,
      "eval_runtime": 146.2563,
      "eval_samples_per_second": 2.345,
      "eval_steps_per_second": 0.294,
      "step": 100
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.4982572012636904e-07,
      "logits/chosen": -2.0055313110351562,
      "logits/rejected": -2.00311541557312,
      "logps/chosen": -33.23518753051758,
      "logps/rejected": -34.03091812133789,
      "loss": 0.3233,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": 0.007764431647956371,
      "rewards/margins": 0.009159245528280735,
      "rewards/rejected": -0.0013948131818324327,
      "step": 110
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.353806263777677e-07,
      "logits/chosen": -2.016556978225708,
      "logits/rejected": -2.0081825256347656,
      "logps/chosen": -32.435211181640625,
      "logps/rejected": -32.18501281738281,
      "loss": 0.3121,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.00872363243252039,
      "rewards/margins": 0.029877260327339172,
      "rewards/rejected": -0.02115362510085106,
      "step": 120
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.194082707715275e-07,
      "logits/chosen": -2.0461266040802,
      "logits/rejected": -2.03806734085083,
      "logps/chosen": -30.502710342407227,
      "logps/rejected": -32.0546760559082,
      "loss": 0.3674,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": -0.02429451420903206,
      "rewards/margins": -0.02279272861778736,
      "rewards/rejected": -0.0015017850091680884,
      "step": 130
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.020402418666621e-07,
      "logits/chosen": -1.9767634868621826,
      "logits/rejected": -1.9870494604110718,
      "logps/chosen": -31.402328491210938,
      "logps/rejected": -32.55817794799805,
      "loss": 0.3125,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.015601096674799919,
      "rewards/margins": 0.03657008707523346,
      "rewards/rejected": -0.02096899226307869,
      "step": 140
    },
    {
      "epoch": 0.39,
      "learning_rate": 3.8341962650351185e-07,
      "logits/chosen": -1.8908436298370361,
      "logits/rejected": -1.8919413089752197,
      "logps/chosen": -34.207401275634766,
      "logps/rejected": -34.783695220947266,
      "loss": 0.3373,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": -0.013235817663371563,
      "rewards/margins": 0.01055879332125187,
      "rewards/rejected": -0.023794615641236305,
      "step": 150
    },
    {
      "epoch": 0.42,
      "learning_rate": 3.636998309800572e-07,
      "logits/chosen": -1.942582368850708,
      "logits/rejected": -1.9391053915023804,
      "logps/chosen": -36.15548324584961,
      "logps/rejected": -32.729759216308594,
      "loss": 0.3207,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.014574882574379444,
      "rewards/margins": 0.020880455151200294,
      "rewards/rejected": -0.006305573042482138,
      "step": 160
    },
    {
      "epoch": 0.44,
      "learning_rate": 3.430433172111807e-07,
      "logits/chosen": -2.042206287384033,
      "logits/rejected": -2.034825086593628,
      "logps/chosen": -33.77631378173828,
      "logps/rejected": -31.35869789123535,
      "loss": 0.3201,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.019394155591726303,
      "rewards/margins": 0.025556519627571106,
      "rewards/rejected": -0.00616236450150609,
      "step": 170
    },
    {
      "epoch": 0.47,
      "learning_rate": 3.216202642830543e-07,
      "logits/chosen": -2.0474154949188232,
      "logits/rejected": -2.0526936054229736,
      "logps/chosen": -32.521339416503906,
      "logps/rejected": -32.497581481933594,
      "loss": 0.2999,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.021103620529174805,
      "rewards/margins": 0.032301973551511765,
      "rewards/rejected": -0.01119835115969181,
      "step": 180
    },
    {
      "epoch": 0.49,
      "learning_rate": 2.9960716642946403e-07,
      "logits/chosen": -2.0485098361968994,
      "logits/rejected": -2.0457375049591064,
      "logps/chosen": -31.49017906188965,
      "logps/rejected": -31.345867156982422,
      "loss": 0.3315,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.001104071387089789,
      "rewards/margins": 0.023781923577189445,
      "rewards/rejected": -0.02488599717617035,
      "step": 190
    },
    {
      "epoch": 0.52,
      "learning_rate": 2.771853789806683e-07,
      "logits/chosen": -1.9183101654052734,
      "logits/rejected": -1.9229921102523804,
      "logps/chosen": -31.60622215270996,
      "logps/rejected": -32.805049896240234,
      "loss": 0.3306,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 5.159676220500842e-05,
      "rewards/margins": 0.012757277116179466,
      "rewards/rejected": -0.012705676257610321,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": -2.2434937953948975,
      "eval_logits/rejected": -2.2386233806610107,
      "eval_logps/chosen": -34.02851867675781,
      "eval_logps/rejected": -37.51131057739258,
      "eval_loss": 0.3466452658176422,
      "eval_rewards/accuracies": 0.5020764470100403,
      "eval_rewards/chosen": 0.005428452976047993,
      "eval_rewards/margins": 0.0006523749325424433,
      "eval_rewards/rejected": 0.00477607874199748,
      "eval_runtime": 145.8217,
      "eval_samples_per_second": 2.352,
      "eval_steps_per_second": 0.295,
      "step": 200
    },
    {
      "epoch": 0.55,
      "learning_rate": 2.5453962426402e-07,
      "logits/chosen": -2.032118558883667,
      "logits/rejected": -2.042782783508301,
      "logps/chosen": -31.939167022705078,
      "logps/rejected": -33.88642501831055,
      "loss": 0.2871,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.02892182394862175,
      "rewards/margins": 0.050276655703783035,
      "rewards/rejected": -0.021354828029870987,
      "step": 210
    },
    {
      "epoch": 0.57,
      "learning_rate": 2.318564697655179e-07,
      "logits/chosen": -1.925293207168579,
      "logits/rejected": -1.9401395320892334,
      "logps/chosen": -30.106586456298828,
      "logps/rejected": -31.56791114807129,
      "loss": 0.3327,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.0044566914439201355,
      "rewards/margins": 0.013218318112194538,
      "rewards/rejected": -0.008761629462242126,
      "step": 220
    },
    {
      "epoch": 0.6,
      "learning_rate": 2.093227910899832e-07,
      "logits/chosen": -1.9831531047821045,
      "logits/rejected": -1.9871160984039307,
      "logps/chosen": -33.386024475097656,
      "logps/rejected": -31.5727481842041,
      "loss": 0.2974,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.028298119083046913,
      "rewards/margins": 0.04939670115709305,
      "rewards/rejected": -0.021098580211400986,
      "step": 230
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.8712423238279356e-07,
      "logits/chosen": -1.9833450317382812,
      "logits/rejected": -1.9613698720932007,
      "logps/chosen": -34.1508903503418,
      "logps/rejected": -34.95817947387695,
      "loss": 0.3388,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.002569653559476137,
      "rewards/margins": 0.00921549927443266,
      "rewards/rejected": -0.006645847111940384,
      "step": 240
    },
    {
      "epoch": 0.65,
      "learning_rate": 1.654436768970182e-07,
      "logits/chosen": -2.0248680114746094,
      "logits/rejected": -2.021559238433838,
      "logps/chosen": -32.907005310058594,
      "logps/rejected": -36.22737121582031,
      "loss": 0.3221,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.016254058107733727,
      "rewards/margins": 0.018458303064107895,
      "rewards/rejected": -0.002204245189204812,
      "step": 250
    },
    {
      "epoch": 0.68,
      "learning_rate": 1.444597403062196e-07,
      "logits/chosen": -1.8913663625717163,
      "logits/rejected": -1.8889217376708984,
      "logps/chosen": -34.18629455566406,
      "logps/rejected": -35.490962982177734,
      "loss": 0.3601,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.0018952053505927324,
      "rewards/margins": -0.010397391393780708,
      "rewards/rejected": 0.012292595580220222,
      "step": 260
    },
    {
      "epoch": 0.7,
      "learning_rate": 1.2434529917578887e-07,
      "logits/chosen": -1.8756721019744873,
      "logits/rejected": -1.8731330633163452,
      "logps/chosen": -34.39625549316406,
      "logps/rejected": -31.76236343383789,
      "loss": 0.3492,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": 0.001211274997331202,
      "rewards/margins": -0.006859898567199707,
      "rewards/rejected": 0.008071173913776875,
      "step": 270
    },
    {
      "epoch": 0.73,
      "learning_rate": 1.0526606671603521e-07,
      "logits/chosen": -1.9795773029327393,
      "logits/rejected": -1.9689537286758423,
      "logps/chosen": -35.320404052734375,
      "logps/rejected": -31.86062240600586,
      "loss": 0.298,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.032611697912216187,
      "rewards/margins": 0.05027870461344719,
      "rewards/rejected": -0.017666997388005257,
      "step": 280
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.737922755071453e-08,
      "logits/chosen": -2.075582981109619,
      "logits/rejected": -2.0605711936950684,
      "logps/chosen": -30.907567977905273,
      "logps/rejected": -32.65256881713867,
      "loss": 0.3186,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": 0.014904007315635681,
      "rewards/margins": 0.015564213506877422,
      "rewards/rejected": -0.0006602049106732011,
      "step": 290
    },
    {
      "epoch": 0.78,
      "learning_rate": 7.08321427484816e-08,
      "logits/chosen": -1.9468443393707275,
      "logits/rejected": -1.944310188293457,
      "logps/chosen": -32.90850067138672,
      "logps/rejected": -30.801532745361328,
      "loss": 0.3327,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": 0.016746358945965767,
      "rewards/margins": 0.002998964861035347,
      "rewards/rejected": 0.01374739408493042,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": -2.243875741958618,
      "eval_logits/rejected": -2.238999366760254,
      "eval_logps/chosen": -34.01763916015625,
      "eval_logps/rejected": -37.51422882080078,
      "eval_loss": 0.32764142751693726,
      "eval_rewards/accuracies": 0.5365448594093323,
      "eval_rewards/chosen": 0.015219501219689846,
      "eval_rewards/margins": 0.01306714303791523,
      "eval_rewards/rejected": 0.0021523565519601107,
      "eval_runtime": 145.9076,
      "eval_samples_per_second": 2.351,
      "eval_steps_per_second": 0.295,
      "step": 300
    },
    {
      "epoch": 0.81,
      "learning_rate": 5.576113578589034e-08,
      "logits/chosen": -1.9285104274749756,
      "logits/rejected": -1.925238847732544,
      "logps/chosen": -31.57692527770996,
      "logps/rejected": -33.744144439697266,
      "loss": 0.3143,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.023783722892403603,
      "rewards/margins": 0.025577496737241745,
      "rewards/rejected": -0.0017937703523784876,
      "step": 310
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.229036944380912e-08,
      "logits/chosen": -1.980386734008789,
      "logits/rejected": -1.9680719375610352,
      "logps/chosen": -34.57415771484375,
      "logps/rejected": -33.55961227416992,
      "loss": 0.31,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.01972355879843235,
      "rewards/margins": 0.03841744735836983,
      "rewards/rejected": -0.018693890422582626,
      "step": 320
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.053082288996112e-08,
      "logits/chosen": -2.015927314758301,
      "logits/rejected": -2.014465570449829,
      "logps/chosen": -33.47549057006836,
      "logps/rejected": -32.47983169555664,
      "loss": 0.3454,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": 0.009270393289625645,
      "rewards/margins": 0.009491671808063984,
      "rewards/rejected": -0.0002212822437286377,
      "step": 330
    },
    {
      "epoch": 0.88,
      "learning_rate": 2.05793773749158e-08,
      "logits/chosen": -2.102839946746826,
      "logits/rejected": -2.0870471000671387,
      "logps/chosen": -34.18696212768555,
      "logps/rejected": -33.075828552246094,
      "loss": 0.4092,
      "rewards/accuracies": 0.4124999940395355,
      "rewards/chosen": -0.006814646068960428,
      "rewards/margins": -0.04479615017771721,
      "rewards/rejected": 0.03798150271177292,
      "step": 340
    },
    {
      "epoch": 0.91,
      "learning_rate": 1.251801807404168e-08,
      "logits/chosen": -1.9747005701065063,
      "logits/rejected": -1.9737575054168701,
      "logps/chosen": -33.231842041015625,
      "logps/rejected": -32.461700439453125,
      "loss": 0.3207,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.04523162543773651,
      "rewards/margins": 0.04212886095046997,
      "rewards/rejected": 0.003102770773693919,
      "step": 350
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.41315865106129e-09,
      "logits/chosen": -1.9306046962738037,
      "logits/rejected": -1.9409637451171875,
      "logps/chosen": -32.217220306396484,
      "logps/rejected": -35.306907653808594,
      "loss": 0.3411,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": -0.011275654658675194,
      "rewards/margins": 0.005062489304691553,
      "rewards/rejected": -0.01633814536035061,
      "step": 360
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.3150941078050324e-09,
      "logits/chosen": -2.0694775581359863,
      "logits/rejected": -2.0629212856292725,
      "logps/chosen": -33.65714645385742,
      "logps/rejected": -29.221160888671875,
      "loss": 0.3549,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": -0.004660401493310928,
      "rewards/margins": -0.009401795454323292,
      "rewards/rejected": 0.0047413939610123634,
      "step": 370
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.575864278703266e-10,
      "logits/chosen": -1.9296157360076904,
      "logits/rejected": -1.931777000427246,
      "logps/chosen": -34.22825241088867,
      "logps/rejected": -30.912120819091797,
      "loss": 0.2952,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.009714365005493164,
      "rewards/margins": 0.04525749385356903,
      "rewards/rejected": -0.035543132573366165,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.0744757689438857,
      "train_runtime": 628.0651,
      "train_samples_per_second": 4.902,
      "train_steps_per_second": 0.613
    }
  ],
  "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
}