{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9984301412872841,
  "eval_steps": 500,
  "global_step": 159,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 3.125e-08,
      "logits/chosen": -2.9105112552642822,
      "logits/rejected": -2.8928871154785156,
      "logps/chosen": -301.8572082519531,
      "logps/rejected": -318.947998046875,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.06,
      "learning_rate": 3.1249999999999997e-07,
      "logits/chosen": -2.8155131340026855,
      "logits/rejected": -2.831986904144287,
      "logps/chosen": -259.8145751953125,
      "logps/rejected": -244.01991271972656,
      "loss": 0.693,
      "rewards/accuracies": 0.4791666567325592,
      "rewards/chosen": -0.00019058177713304758,
      "rewards/margins": 0.0005300731863826513,
      "rewards/rejected": -0.0007206548471003771,
      "step": 10
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.990353313429303e-07,
      "logits/chosen": -2.8561551570892334,
      "logits/rejected": -2.8609530925750732,
      "logps/chosen": -251.63162231445312,
      "logps/rejected": -243.1693878173828,
      "loss": 0.6879,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.005067775025963783,
      "rewards/margins": 0.011929959990084171,
      "rewards/rejected": -0.006862184498459101,
      "step": 20
    },
    {
      "epoch": 0.19,
      "learning_rate": 4.882681251368548e-07,
      "logits/chosen": -2.7307772636413574,
      "logits/rejected": -2.7042434215545654,
      "logps/chosen": -256.0709228515625,
      "logps/rejected": -243.78384399414062,
      "loss": 0.6719,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": 0.02920190431177616,
      "rewards/margins": 0.05106853321194649,
      "rewards/rejected": -0.021866630762815475,
      "step": 30
    },
    {
      "epoch": 0.25,
      "learning_rate": 4.6604720940421207e-07,
      "logits/chosen": -2.812223196029663,
      "logits/rejected": -2.7959158420562744,
      "logps/chosen": -304.9759826660156,
      "logps/rejected": -276.5176086425781,
      "loss": 0.6489,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.030368661507964134,
      "rewards/margins": 0.09863166511058807,
      "rewards/rejected": -0.12900032103061676,
      "step": 40
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.3344075855595097e-07,
      "logits/chosen": -2.78918194770813,
      "logits/rejected": -2.7687530517578125,
      "logps/chosen": -269.62628173828125,
      "logps/rejected": -265.57623291015625,
      "loss": 0.6335,
      "rewards/accuracies": 0.6812499761581421,
      "rewards/chosen": -0.12396706640720367,
      "rewards/margins": 0.1893375813961029,
      "rewards/rejected": -0.3133046329021454,
      "step": 50
    },
    {
      "epoch": 0.38,
      "learning_rate": 3.920161866827889e-07,
      "logits/chosen": -2.7410378456115723,
      "logits/rejected": -2.738602876663208,
      "logps/chosen": -276.0849304199219,
      "logps/rejected": -306.392333984375,
      "loss": 0.6141,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.26490938663482666,
      "rewards/margins": 0.2907998859882355,
      "rewards/rejected": -0.5557092428207397,
      "step": 60
    },
    {
      "epoch": 0.44,
      "learning_rate": 3.4376480090239047e-07,
      "logits/chosen": -2.7223474979400635,
      "logits/rejected": -2.7086873054504395,
      "logps/chosen": -273.6649475097656,
      "logps/rejected": -312.43463134765625,
      "loss": 0.5807,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.42463892698287964,
      "rewards/margins": 0.36430081725120544,
      "rewards/rejected": -0.7889397740364075,
      "step": 70
    },
    {
      "epoch": 0.5,
      "learning_rate": 2.910060778827554e-07,
      "logits/chosen": -2.834973096847534,
      "logits/rejected": -2.783313274383545,
      "logps/chosen": -341.79351806640625,
      "logps/rejected": -299.11236572265625,
      "loss": 0.6015,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.3846871852874756,
      "rewards/margins": 0.3705058991909027,
      "rewards/rejected": -0.7551931142807007,
      "step": 80
    },
    {
      "epoch": 0.57,
      "learning_rate": 2.3627616503391812e-07,
      "logits/chosen": -2.8397436141967773,
      "logits/rejected": -2.7767767906188965,
      "logps/chosen": -355.56182861328125,
      "logps/rejected": -376.3609924316406,
      "loss": 0.573,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.4431120753288269,
      "rewards/margins": 0.5175312757492065,
      "rewards/rejected": -0.9606432914733887,
      "step": 90
    },
    {
      "epoch": 0.63,
      "learning_rate": 1.8220596619089573e-07,
      "logits/chosen": -2.7578227519989014,
      "logits/rejected": -2.7007813453674316,
      "logps/chosen": -313.7681884765625,
      "logps/rejected": -321.396240234375,
      "loss": 0.5849,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.5910673141479492,
      "rewards/margins": 0.26297900080680847,
      "rewards/rejected": -0.8540462255477905,
      "step": 100
    },
    {
      "epoch": 0.69,
      "learning_rate": 1.3139467229135998e-07,
      "logits/chosen": -2.7411677837371826,
      "logits/rejected": -2.6979310512542725,
      "logps/chosen": -313.4439392089844,
      "logps/rejected": -339.30499267578125,
      "loss": 0.5614,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": -0.43303579092025757,
      "rewards/margins": 0.34360843896865845,
      "rewards/rejected": -0.7766442894935608,
      "step": 110
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.628481651367875e-08,
      "logits/chosen": -2.738893747329712,
      "logits/rejected": -2.725003480911255,
      "logps/chosen": -334.6244812011719,
      "logps/rejected": -361.8310546875,
      "loss": 0.5709,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.4588940739631653,
      "rewards/margins": 0.5612925291061401,
      "rewards/rejected": -1.0201866626739502,
      "step": 120
    },
    {
      "epoch": 0.82,
      "learning_rate": 4.904486005914027e-08,
      "logits/chosen": -2.654109001159668,
      "logits/rejected": -2.6659202575683594,
      "logps/chosen": -337.5686950683594,
      "logps/rejected": -347.3365173339844,
      "loss": 0.5745,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -0.6903926134109497,
      "rewards/margins": 0.3957400321960449,
      "rewards/rejected": -1.0861326456069946,
      "step": 130
    },
    {
      "epoch": 0.88,
      "learning_rate": 2.1464952759020856e-08,
      "logits/chosen": -2.731912851333618,
      "logits/rejected": -2.693276882171631,
      "logps/chosen": -345.1060485839844,
      "logps/rejected": -374.956298828125,
      "loss": 0.5394,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.6538197994232178,
      "rewards/margins": 0.5970135927200317,
      "rewards/rejected": -1.25083327293396,
      "step": 140
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.8708793644441086e-09,
      "logits/chosen": -2.6765170097351074,
      "logits/rejected": -2.6645171642303467,
      "logps/chosen": -338.6192932128906,
      "logps/rejected": -356.35687255859375,
      "loss": 0.5494,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.6419016122817993,
      "rewards/margins": 0.5184969305992126,
      "rewards/rejected": -1.1603987216949463,
      "step": 150
    },
    {
      "epoch": 1.0,
      "step": 159,
      "total_flos": 0.0,
      "train_loss": 0.6025606791178385,
      "train_runtime": 2664.2013,
      "train_samples_per_second": 7.649,
      "train_steps_per_second": 0.06
    }
  ],
  "logging_steps": 10,
  "max_steps": 159,
  "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
}