{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.32666394446712943,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.016333197223356473,
      "grad_norm": 1.1593419313430786,
      "learning_rate": 4.9996340229869763e-05,
      "logits/chosen": -1.0346699953079224,
      "logits/rejected": -2.6695446968078613,
      "logps/chosen": -1.101169466972351,
      "logps/rejected": -1.0235556364059448,
      "loss": 1.178,
      "odds_ratio_loss": 0.768240213394165,
      "rewards/accuracies": 0.3499999940395355,
      "rewards/chosen": -0.11011694371700287,
      "rewards/margins": -0.007761377841234207,
      "rewards/rejected": -0.10235557705163956,
      "sft_loss": 1.101169466972351,
      "step": 5
    },
    {
      "epoch": 0.032666394446712946,
      "grad_norm": 1.1402696371078491,
      "learning_rate": 4.9985361990992455e-05,
      "logits/chosen": -0.8193758726119995,
      "logits/rejected": -2.842833995819092,
      "logps/chosen": -1.0217273235321045,
      "logps/rejected": -1.0100481510162354,
      "loss": 1.0927,
      "odds_ratio_loss": 0.7092453837394714,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.10217274725437164,
      "rewards/margins": -0.0011679243762046099,
      "rewards/rejected": -0.10100481659173965,
      "sft_loss": 1.0217273235321045,
      "step": 10
    },
    {
      "epoch": 0.04899959167006942,
      "grad_norm": 1.3554140329360962,
      "learning_rate": 4.996706849759453e-05,
      "logits/chosen": -0.7696830034255981,
      "logits/rejected": -2.703421115875244,
      "logps/chosen": -1.0294106006622314,
      "logps/rejected": -1.0599125623703003,
      "loss": 1.0993,
      "odds_ratio_loss": 0.6984036564826965,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.10294107347726822,
      "rewards/margins": 0.003050185739994049,
      "rewards/rejected": -0.10599125921726227,
      "sft_loss": 1.0294106006622314,
      "step": 15
    },
    {
      "epoch": 0.06533278889342589,
      "grad_norm": 1.406853199005127,
      "learning_rate": 4.9941465105674435e-05,
      "logits/chosen": -0.698488712310791,
      "logits/rejected": -2.808516263961792,
      "logps/chosen": -1.0651928186416626,
      "logps/rejected": -0.996777355670929,
      "loss": 1.1402,
      "odds_ratio_loss": 0.7504132390022278,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": -0.10651928186416626,
      "rewards/margins": -0.006841552909463644,
      "rewards/rejected": -0.09967772662639618,
      "sft_loss": 1.0651928186416626,
      "step": 20
    },
    {
      "epoch": 0.08166598611678236,
      "grad_norm": 1.238513469696045,
      "learning_rate": 4.99085593114345e-05,
      "logits/chosen": -0.6306881904602051,
      "logits/rejected": -2.698362350463867,
      "logps/chosen": -1.0242538452148438,
      "logps/rejected": -1.0320734977722168,
      "loss": 1.0937,
      "odds_ratio_loss": 0.6945221424102783,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.10242538154125214,
      "rewards/margins": 0.0007819652673788369,
      "rewards/rejected": -0.10320734977722168,
      "sft_loss": 1.0242538452148438,
      "step": 25
    },
    {
      "epoch": 0.09799918334013884,
      "grad_norm": 1.9588699340820312,
      "learning_rate": 4.986836074908616e-05,
      "logits/chosen": -0.5534936189651489,
      "logits/rejected": -2.678889036178589,
      "logps/chosen": -1.1003000736236572,
      "logps/rejected": -1.0425399541854858,
      "loss": 1.1745,
      "odds_ratio_loss": 0.7416313290596008,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.11003003269433975,
      "rewards/margins": -0.005776026751846075,
      "rewards/rejected": -0.10425399243831635,
      "sft_loss": 1.1003000736236572,
      "step": 30
    },
    {
      "epoch": 0.1143323805634953,
      "grad_norm": 1.1857821941375732,
      "learning_rate": 4.982088118802922e-05,
      "logits/chosen": -0.6129628419876099,
      "logits/rejected": -2.876364231109619,
      "logps/chosen": -0.9628928303718567,
      "logps/rejected": -0.952060341835022,
      "loss": 1.0336,
      "odds_ratio_loss": 0.7068274617195129,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": -0.09628929197788239,
      "rewards/margins": -0.001083258306607604,
      "rewards/rejected": -0.09520602226257324,
      "sft_loss": 0.9628928303718567,
      "step": 35
    },
    {
      "epoch": 0.13066557778685178,
      "grad_norm": 1.33924400806427,
      "learning_rate": 4.976613452940604e-05,
      "logits/chosen": -0.7091814279556274,
      "logits/rejected": -2.6595664024353027,
      "logps/chosen": -0.89775550365448,
      "logps/rejected": -0.935952365398407,
      "loss": 0.9642,
      "odds_ratio_loss": 0.6644185781478882,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.08977555483579636,
      "rewards/margins": 0.003819683101028204,
      "rewards/rejected": -0.0935952290892601,
      "sft_loss": 0.89775550365448,
      "step": 40
    },
    {
      "epoch": 0.14699877501020825,
      "grad_norm": 1.136667013168335,
      "learning_rate": 4.9704136802031485e-05,
      "logits/chosen": -0.7233263850212097,
      "logits/rejected": -2.6673388481140137,
      "logps/chosen": -0.9036429524421692,
      "logps/rejected": -0.9419530630111694,
      "loss": 0.9699,
      "odds_ratio_loss": 0.662254273891449,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.09036429226398468,
      "rewards/margins": 0.0038310133386403322,
      "rewards/rejected": -0.09419531375169754,
      "sft_loss": 0.9036429524421692,
      "step": 45
    },
    {
      "epoch": 0.16333197223356472,
      "grad_norm": 1.4290399551391602,
      "learning_rate": 4.9634906157700036e-05,
      "logits/chosen": -0.6113058924674988,
      "logits/rejected": -2.705373764038086,
      "logps/chosen": -0.9550830125808716,
      "logps/rejected": -0.9817631840705872,
      "loss": 1.0223,
      "odds_ratio_loss": 0.6725824475288391,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": -0.09550830721855164,
      "rewards/margins": 0.002668013097718358,
      "rewards/rejected": -0.09817632287740707,
      "sft_loss": 0.9550830125808716,
      "step": 50
    },
    {
      "epoch": 0.17966516945692118,
      "grad_norm": 1.4945589303970337,
      "learning_rate": 4.955846286587122e-05,
      "logits/chosen": -0.647710919380188,
      "logits/rejected": -2.7208316326141357,
      "logps/chosen": -0.9462896585464478,
      "logps/rejected": -0.9507497549057007,
      "loss": 1.0148,
      "odds_ratio_loss": 0.6852728128433228,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": -0.0946289673447609,
      "rewards/margins": 0.0004460157942958176,
      "rewards/rejected": -0.09507497400045395,
      "sft_loss": 0.9462896585464478,
      "step": 55
    },
    {
      "epoch": 0.19599836668027767,
      "grad_norm": 1.3553156852722168,
      "learning_rate": 4.9474829307735115e-05,
      "logits/chosen": -0.6279743909835815,
      "logits/rejected": -2.6854496002197266,
      "logps/chosen": -0.8896121978759766,
      "logps/rejected": -0.9993942975997925,
      "loss": 0.9509,
      "odds_ratio_loss": 0.6129180192947388,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.08896122127771378,
      "rewards/margins": 0.010978218168020248,
      "rewards/rejected": -0.09993942826986313,
      "sft_loss": 0.8896121978759766,
      "step": 60
    },
    {
      "epoch": 0.21233156390363414,
      "grad_norm": 1.4247411489486694,
      "learning_rate": 4.9384029969659573e-05,
      "logits/chosen": -0.698381781578064,
      "logits/rejected": -2.7967753410339355,
      "logps/chosen": -0.9455701112747192,
      "logps/rejected": -0.9564310908317566,
      "loss": 1.0147,
      "odds_ratio_loss": 0.6911824941635132,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.0945570170879364,
      "rewards/margins": 0.0010860899928957224,
      "rewards/rejected": -0.09564311057329178,
      "sft_loss": 0.9455701112747192,
      "step": 65
    },
    {
      "epoch": 0.2286647611269906,
      "grad_norm": 1.6154508590698242,
      "learning_rate": 4.9286091436021015e-05,
      "logits/chosen": -0.595032274723053,
      "logits/rejected": -2.740784168243408,
      "logps/chosen": -0.9214372634887695,
      "logps/rejected": -1.0210167169570923,
      "loss": 0.9836,
      "odds_ratio_loss": 0.6212525963783264,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.09214373677968979,
      "rewards/margins": 0.00995794590562582,
      "rewards/rejected": -0.10210166871547699,
      "sft_loss": 0.9214372634887695,
      "step": 70
    },
    {
      "epoch": 0.24499795835034707,
      "grad_norm": 1.8231102228164673,
      "learning_rate": 4.918104238142104e-05,
      "logits/chosen": -0.5186889171600342,
      "logits/rejected": -2.7576358318328857,
      "logps/chosen": -0.9789311289787292,
      "logps/rejected": -1.0169849395751953,
      "loss": 1.0454,
      "odds_ratio_loss": 0.6644267439842224,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -0.09789310395717621,
      "rewards/margins": 0.0038053765892982483,
      "rewards/rejected": -0.10169849544763565,
      "sft_loss": 0.9789311289787292,
      "step": 75
    },
    {
      "epoch": 0.26133115557370357,
      "grad_norm": 2.05936336517334,
      "learning_rate": 4.906891356229103e-05,
      "logits/chosen": -0.47254711389541626,
      "logits/rejected": -2.68280029296875,
      "logps/chosen": -0.9620391726493835,
      "logps/rejected": -1.0026928186416626,
      "loss": 1.0294,
      "odds_ratio_loss": 0.6731926202774048,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.09620390832424164,
      "rewards/margins": 0.004065364599227905,
      "rewards/rejected": -0.10026928037405014,
      "sft_loss": 0.9620391726493835,
      "step": 80
    },
    {
      "epoch": 0.27766435279706003,
      "grad_norm": 1.6302626132965088,
      "learning_rate": 4.894973780788722e-05,
      "logits/chosen": -0.5315309762954712,
      "logits/rejected": -2.7227425575256348,
      "logps/chosen": -0.9423310160636902,
      "logps/rejected": -1.071781039237976,
      "loss": 1.0032,
      "odds_ratio_loss": 0.6090766191482544,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.09423311054706573,
      "rewards/margins": 0.012944999150931835,
      "rewards/rejected": -0.10717810690402985,
      "sft_loss": 0.9423310160636902,
      "step": 85
    },
    {
      "epoch": 0.2939975500204165,
      "grad_norm": 1.7891793251037598,
      "learning_rate": 4.882355001067892e-05,
      "logits/chosen": -0.5275685787200928,
      "logits/rejected": -2.6127567291259766,
      "logps/chosen": -0.9351356625556946,
      "logps/rejected": -1.0083874464035034,
      "loss": 0.9995,
      "odds_ratio_loss": 0.6435801386833191,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.09351354837417603,
      "rewards/margins": 0.0073251971043646336,
      "rewards/rejected": -0.10083875805139542,
      "sft_loss": 0.9351356625556946,
      "step": 90
    },
    {
      "epoch": 0.31033074724377296,
      "grad_norm": 1.852216124534607,
      "learning_rate": 4.869038711613259e-05,
      "logits/chosen": -0.5760647058486938,
      "logits/rejected": -2.722364664077759,
      "logps/chosen": -0.9301847219467163,
      "logps/rejected": -1.0749822854995728,
      "loss": 0.9903,
      "odds_ratio_loss": 0.601297914981842,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.09301847964525223,
      "rewards/margins": 0.014479748904705048,
      "rewards/rejected": -0.10749821364879608,
      "sft_loss": 0.9301847219467163,
      "step": 95
    },
    {
      "epoch": 0.32666394446712943,
      "grad_norm": 1.8597307205200195,
      "learning_rate": 4.855028811189496e-05,
      "logits/chosen": -0.5871164202690125,
      "logits/rejected": -2.6969285011291504,
      "logps/chosen": -1.0701930522918701,
      "logps/rejected": -1.0556358098983765,
      "loss": 1.1427,
      "odds_ratio_loss": 0.7247555255889893,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.10701930522918701,
      "rewards/margins": -0.0014557272661477327,
      "rewards/rejected": -0.10556356608867645,
      "sft_loss": 1.0701930522918701,
      "step": 100
    }
  ],
  "logging_steps": 5,
  "max_steps": 918,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "total_flos": 4.118738025735782e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}