{
  "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": 104.5,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": 88.18099975585938,
      "logits/rejected": 88.25153350830078,
      "logps/chosen": -29.073104858398438,
      "logps/rejected": -26.25731658935547,
      "loss": 1.0,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 78.0,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": 81.08700561523438,
      "logits/rejected": 80.79035186767578,
      "logps/chosen": -34.29237365722656,
      "logps/rejected": -33.04549026489258,
      "loss": 0.9866,
      "rewards/accuracies": 0.3888888955116272,
      "rewards/chosen": -0.040518708527088165,
      "rewards/margins": 0.029149238020181656,
      "rewards/rejected": -0.06966794282197952,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 85.5,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": 80.68280029296875,
      "logits/rejected": 80.57057189941406,
      "logps/chosen": -33.53000259399414,
      "logps/rejected": -30.8519287109375,
      "loss": 0.8984,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.12792269885540009,
      "rewards/margins": 0.18747423589229584,
      "rewards/rejected": -0.05955154448747635,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 73.0,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": 82.52570343017578,
      "logits/rejected": 82.552734375,
      "logps/chosen": -33.96383285522461,
      "logps/rejected": -31.202930450439453,
      "loss": 1.1955,
      "rewards/accuracies": 0.4000000059604645,
      "rewards/chosen": 0.03918968886137009,
      "rewards/margins": -0.1375342458486557,
      "rewards/rejected": 0.1767239272594452,
      "step": 30
    },
    {
      "epoch": 0.1,
      "grad_norm": 74.5,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": 80.97098541259766,
      "logits/rejected": 80.9694595336914,
      "logps/chosen": -32.8037109375,
      "logps/rejected": -33.232460021972656,
      "loss": 0.9309,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.2962803840637207,
      "rewards/margins": 0.1874258816242218,
      "rewards/rejected": 0.10885453224182129,
      "step": 40
    },
    {
      "epoch": 0.13,
      "grad_norm": 70.0,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": 78.56432342529297,
      "logits/rejected": 78.58203125,
      "logps/chosen": -30.68828773498535,
      "logps/rejected": -30.789108276367188,
      "loss": 0.9855,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.4135704040527344,
      "rewards/margins": 0.18622872233390808,
      "rewards/rejected": 0.2273416966199875,
      "step": 50
    },
    {
      "epoch": 0.16,
      "grad_norm": 84.0,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": 83.11712646484375,
      "logits/rejected": 83.17086791992188,
      "logps/chosen": -30.859710693359375,
      "logps/rejected": -29.577678680419922,
      "loss": 0.9236,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.25236082077026367,
      "rewards/margins": 0.24208447337150574,
      "rewards/rejected": 0.010276327840983868,
      "step": 60
    },
    {
      "epoch": 0.18,
      "grad_norm": 107.0,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": 83.80155944824219,
      "logits/rejected": 83.8315658569336,
      "logps/chosen": -30.606204986572266,
      "logps/rejected": -32.91896057128906,
      "loss": 1.141,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": 0.01825830340385437,
      "rewards/margins": -0.0767388790845871,
      "rewards/rejected": 0.09499720484018326,
      "step": 70
    },
    {
      "epoch": 0.21,
      "grad_norm": 84.5,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": 81.45521545410156,
      "logits/rejected": 81.44440460205078,
      "logps/chosen": -31.48111343383789,
      "logps/rejected": -31.072092056274414,
      "loss": 0.8995,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.01906520128250122,
      "rewards/margins": 0.2706204056739807,
      "rewards/rejected": -0.2515551745891571,
      "step": 80
    },
    {
      "epoch": 0.23,
      "grad_norm": 102.0,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": 78.27713775634766,
      "logits/rejected": 78.24540710449219,
      "logps/chosen": -32.49984359741211,
      "logps/rejected": -31.142908096313477,
      "loss": 0.9467,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.11066800355911255,
      "rewards/margins": 0.2638159692287445,
      "rewards/rejected": -0.15314793586730957,
      "step": 90
    },
    {
      "epoch": 0.26,
      "grad_norm": 73.5,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": 83.56034851074219,
      "logits/rejected": 83.58055877685547,
      "logps/chosen": -34.057029724121094,
      "logps/rejected": -31.78998374938965,
      "loss": 0.9567,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.19836829602718353,
      "rewards/margins": 0.22992074489593506,
      "rewards/rejected": -0.031552452594041824,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": 98.73486328125,
      "eval_logits/rejected": 98.72332763671875,
      "eval_logps/chosen": -32.45964431762695,
      "eval_logps/rejected": -35.88682556152344,
      "eval_loss": 1.1574316024780273,
      "eval_rewards/accuracies": 0.4630398452281952,
      "eval_rewards/chosen": -0.013175476342439651,
      "eval_rewards/margins": -0.07693858444690704,
      "eval_rewards/rejected": 0.0637631043791771,
      "eval_runtime": 104.0187,
      "eval_samples_per_second": 3.297,
      "eval_steps_per_second": 0.413,
      "step": 100
    },
    {
      "epoch": 0.29,
      "grad_norm": 104.0,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": 83.91990661621094,
      "logits/rejected": 83.79493713378906,
      "logps/chosen": -32.23216247558594,
      "logps/rejected": -32.69822692871094,
      "loss": 0.7793,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.6325147151947021,
      "rewards/margins": 0.6663830280303955,
      "rewards/rejected": -0.03386829048395157,
      "step": 110
    },
    {
      "epoch": 0.31,
      "grad_norm": 108.5,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": 84.13631439208984,
      "logits/rejected": 84.24298858642578,
      "logps/chosen": -28.422576904296875,
      "logps/rejected": -35.379127502441406,
      "loss": 0.8738,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.40560561418533325,
      "rewards/margins": 0.34159213304519653,
      "rewards/rejected": 0.06401350349187851,
      "step": 120
    },
    {
      "epoch": 0.34,
      "grad_norm": 55.5,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": 81.52787017822266,
      "logits/rejected": 81.55335998535156,
      "logps/chosen": -30.235469818115234,
      "logps/rejected": -31.976581573486328,
      "loss": 0.8013,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.5375091433525085,
      "rewards/margins": 0.5734738707542419,
      "rewards/rejected": -0.03596482425928116,
      "step": 130
    },
    {
      "epoch": 0.36,
      "grad_norm": 57.75,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": 82.77738189697266,
      "logits/rejected": 82.78349304199219,
      "logps/chosen": -27.04022216796875,
      "logps/rejected": -32.86003875732422,
      "loss": 0.7387,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.32568174600601196,
      "rewards/margins": 0.6750079393386841,
      "rewards/rejected": -0.34932616353034973,
      "step": 140
    },
    {
      "epoch": 0.39,
      "grad_norm": 63.0,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": 81.38068389892578,
      "logits/rejected": 81.35902404785156,
      "logps/chosen": -28.90557289123535,
      "logps/rejected": -32.89466094970703,
      "loss": 0.6488,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.38682785630226135,
      "rewards/margins": 0.6493567228317261,
      "rewards/rejected": -0.26252883672714233,
      "step": 150
    },
    {
      "epoch": 0.42,
      "grad_norm": 66.0,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": 83.2899169921875,
      "logits/rejected": 83.30531311035156,
      "logps/chosen": -33.436553955078125,
      "logps/rejected": -30.43096351623535,
      "loss": 0.6075,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.5601544976234436,
      "rewards/margins": 0.9428972005844116,
      "rewards/rejected": -0.382742702960968,
      "step": 160
    },
    {
      "epoch": 0.44,
      "grad_norm": 70.5,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": 84.12855529785156,
      "logits/rejected": 84.0693359375,
      "logps/chosen": -30.63346290588379,
      "logps/rejected": -32.29146194458008,
      "loss": 0.7575,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.5006591081619263,
      "rewards/margins": 0.717779278755188,
      "rewards/rejected": -0.21712002158164978,
      "step": 170
    },
    {
      "epoch": 0.47,
      "grad_norm": 65.0,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": 81.81861114501953,
      "logits/rejected": 81.79876708984375,
      "logps/chosen": -30.488971710205078,
      "logps/rejected": -31.519611358642578,
      "loss": 0.6265,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.5665422677993774,
      "rewards/margins": 0.8564668893814087,
      "rewards/rejected": -0.289924681186676,
      "step": 180
    },
    {
      "epoch": 0.49,
      "grad_norm": 34.0,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": 83.52326965332031,
      "logits/rejected": 83.50785064697266,
      "logps/chosen": -30.285634994506836,
      "logps/rejected": -30.630422592163086,
      "loss": 0.8287,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.38462626934051514,
      "rewards/margins": 0.5145466327667236,
      "rewards/rejected": -0.12992039322853088,
      "step": 190
    },
    {
      "epoch": 0.52,
      "grad_norm": 50.25,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": 79.0837173461914,
      "logits/rejected": 79.03185272216797,
      "logps/chosen": -33.81468963623047,
      "logps/rejected": -32.4266357421875,
      "loss": 0.8098,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.7504364252090454,
      "rewards/margins": 0.6970199346542358,
      "rewards/rejected": 0.053416453301906586,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": 98.86815643310547,
      "eval_logits/rejected": 98.84379577636719,
      "eval_logps/chosen": -32.561100006103516,
      "eval_logps/rejected": -36.16287612915039,
      "eval_loss": 1.0544954538345337,
      "eval_rewards/accuracies": 0.5278239250183105,
      "eval_rewards/chosen": -0.09434036910533905,
      "eval_rewards/margins": 0.06273789703845978,
      "eval_rewards/rejected": -0.15707828104496002,
      "eval_runtime": 103.8402,
      "eval_samples_per_second": 3.303,
      "eval_steps_per_second": 0.414,
      "step": 200
    },
    {
      "epoch": 0.55,
      "grad_norm": 104.0,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": 81.58848571777344,
      "logits/rejected": 81.49668884277344,
      "logps/chosen": -33.169471740722656,
      "logps/rejected": -34.991939544677734,
      "loss": 0.7585,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.620996356010437,
      "rewards/margins": 0.6256735324859619,
      "rewards/rejected": -0.004677181597799063,
      "step": 210
    },
    {
      "epoch": 0.57,
      "grad_norm": 35.0,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": 83.66175079345703,
      "logits/rejected": 83.74690246582031,
      "logps/chosen": -31.094005584716797,
      "logps/rejected": -31.061717987060547,
      "loss": 0.626,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": 0.683959424495697,
      "rewards/margins": 0.9501722455024719,
      "rewards/rejected": -0.2662127912044525,
      "step": 220
    },
    {
      "epoch": 0.6,
      "grad_norm": 69.5,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": 80.7747802734375,
      "logits/rejected": 80.82144165039062,
      "logps/chosen": -32.322105407714844,
      "logps/rejected": -34.30824661254883,
      "loss": 0.7928,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.38691413402557373,
      "rewards/margins": 0.6237919926643372,
      "rewards/rejected": -0.23687779903411865,
      "step": 230
    },
    {
      "epoch": 0.62,
      "grad_norm": 89.5,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": 83.18328094482422,
      "logits/rejected": 83.4581069946289,
      "logps/chosen": -30.6195125579834,
      "logps/rejected": -31.637020111083984,
      "loss": 0.5475,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.854840099811554,
      "rewards/margins": 0.9388192892074585,
      "rewards/rejected": -0.08397923409938812,
      "step": 240
    },
    {
      "epoch": 0.65,
      "grad_norm": 84.5,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": 81.9409408569336,
      "logits/rejected": 82.00200653076172,
      "logps/chosen": -26.908878326416016,
      "logps/rejected": -30.06099510192871,
      "loss": 0.8063,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.5496150255203247,
      "rewards/margins": 0.6778196692466736,
      "rewards/rejected": -0.12820473313331604,
      "step": 250
    },
    {
      "epoch": 0.68,
      "grad_norm": 60.25,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": 79.24473571777344,
      "logits/rejected": 79.37559509277344,
      "logps/chosen": -30.283336639404297,
      "logps/rejected": -36.4005241394043,
      "loss": 0.4689,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": 1.0221258401870728,
      "rewards/margins": 1.2612489461898804,
      "rewards/rejected": -0.23912319540977478,
      "step": 260
    },
    {
      "epoch": 0.7,
      "grad_norm": 31.875,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": 78.6611328125,
      "logits/rejected": 78.69209289550781,
      "logps/chosen": -30.992252349853516,
      "logps/rejected": -31.770212173461914,
      "loss": 0.6471,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.6293355822563171,
      "rewards/margins": 0.8411690592765808,
      "rewards/rejected": -0.21183356642723083,
      "step": 270
    },
    {
      "epoch": 0.73,
      "grad_norm": 100.0,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": 81.2646255493164,
      "logits/rejected": 81.05351257324219,
      "logps/chosen": -31.10286521911621,
      "logps/rejected": -29.866031646728516,
      "loss": 0.7333,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.5656233429908752,
      "rewards/margins": 0.7248517274856567,
      "rewards/rejected": -0.15922844409942627,
      "step": 280
    },
    {
      "epoch": 0.75,
      "grad_norm": 61.25,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": 81.45980072021484,
      "logits/rejected": 81.37715148925781,
      "logps/chosen": -33.03575134277344,
      "logps/rejected": -32.42512512207031,
      "loss": 0.6231,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.8547972440719604,
      "rewards/margins": 1.1814041137695312,
      "rewards/rejected": -0.32660672068595886,
      "step": 290
    },
    {
      "epoch": 0.78,
      "grad_norm": 78.0,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": 77.12594604492188,
      "logits/rejected": 77.21607971191406,
      "logps/chosen": -32.209590911865234,
      "logps/rejected": -29.23952865600586,
      "loss": 0.6965,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.9568928480148315,
      "rewards/margins": 1.066897988319397,
      "rewards/rejected": -0.11000506579875946,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": 98.90528106689453,
      "eval_logits/rejected": 98.8800277709961,
      "eval_logps/chosen": -32.50230407714844,
      "eval_logps/rejected": -36.192344665527344,
      "eval_loss": 0.9868877530097961,
      "eval_rewards/accuracies": 0.5830564498901367,
      "eval_rewards/chosen": -0.047303199768066406,
      "eval_rewards/margins": 0.1333501785993576,
      "eval_rewards/rejected": -0.180653378367424,
      "eval_runtime": 103.7665,
      "eval_samples_per_second": 3.305,
      "eval_steps_per_second": 0.414,
      "step": 300
    },
    {
      "epoch": 0.81,
      "grad_norm": 59.25,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": 84.09150695800781,
      "logits/rejected": 84.11959075927734,
      "logps/chosen": -30.127155303955078,
      "logps/rejected": -32.29438781738281,
      "loss": 0.7446,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.5891044735908508,
      "rewards/margins": 0.6648039817810059,
      "rewards/rejected": -0.07569954544305801,
      "step": 310
    },
    {
      "epoch": 0.83,
      "grad_norm": 60.25,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": 81.67195892333984,
      "logits/rejected": 81.67171478271484,
      "logps/chosen": -30.566574096679688,
      "logps/rejected": -29.08144187927246,
      "loss": 0.5518,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.8255898356437683,
      "rewards/margins": 0.9551518559455872,
      "rewards/rejected": -0.12956194579601288,
      "step": 320
    },
    {
      "epoch": 0.86,
      "grad_norm": 58.0,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": 78.98292541503906,
      "logits/rejected": 79.03089904785156,
      "logps/chosen": -29.0233211517334,
      "logps/rejected": -32.939231872558594,
      "loss": 0.5283,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 1.063429832458496,
      "rewards/margins": 1.306926965713501,
      "rewards/rejected": -0.24349698424339294,
      "step": 330
    },
    {
      "epoch": 0.88,
      "grad_norm": 88.5,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": 83.09165954589844,
      "logits/rejected": 83.11895751953125,
      "logps/chosen": -32.228084564208984,
      "logps/rejected": -33.656490325927734,
      "loss": 0.6498,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.7872547507286072,
      "rewards/margins": 1.0174915790557861,
      "rewards/rejected": -0.23023685812950134,
      "step": 340
    },
    {
      "epoch": 0.91,
      "grad_norm": 42.5,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": 82.23709869384766,
      "logits/rejected": 82.24420928955078,
      "logps/chosen": -32.544349670410156,
      "logps/rejected": -33.209228515625,
      "loss": 0.6454,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.8795096278190613,
      "rewards/margins": 0.9734523892402649,
      "rewards/rejected": -0.09394274652004242,
      "step": 350
    },
    {
      "epoch": 0.94,
      "grad_norm": 66.0,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": 83.5949935913086,
      "logits/rejected": 83.62043762207031,
      "logps/chosen": -28.472000122070312,
      "logps/rejected": -31.736831665039062,
      "loss": 0.6336,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.8641688227653503,
      "rewards/margins": 0.9337539672851562,
      "rewards/rejected": -0.0695851519703865,
      "step": 360
    },
    {
      "epoch": 0.96,
      "grad_norm": 78.5,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": 83.08488464355469,
      "logits/rejected": 83.10858917236328,
      "logps/chosen": -31.81646156311035,
      "logps/rejected": -35.26220703125,
      "loss": 0.6364,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.847761332988739,
      "rewards/margins": 0.9700264930725098,
      "rewards/rejected": -0.1222650408744812,
      "step": 370
    },
    {
      "epoch": 0.99,
      "grad_norm": 82.5,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": 77.07919311523438,
      "logits/rejected": 76.94573974609375,
      "logps/chosen": -29.726633071899414,
      "logps/rejected": -28.218725204467773,
      "loss": 0.7075,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.6795950531959534,
      "rewards/margins": 0.7261168956756592,
      "rewards/rejected": -0.046521805226802826,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.7616742146479619,
      "train_runtime": 2551.1965,
      "train_samples_per_second": 1.207,
      "train_steps_per_second": 0.151
    }
  ],
  "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
}