{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.971563981042654,
  "eval_steps": 128,
  "global_step": 104,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.018957345971563982,
      "grad_norm": 191.735526981635,
      "learning_rate": 4.545454545454545e-08,
      "logits/chosen": 117.53560638427734,
      "logits/rejected": 126.8960952758789,
      "logps/chosen": -335.40118408203125,
      "logps/rejected": -439.16552734375,
      "loss": 0.8466,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.1895734597156398,
      "grad_norm": 215.27789653097236,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": 135.01699829101562,
      "logits/rejected": 138.35781860351562,
      "logps/chosen": -395.863525390625,
      "logps/rejected": -439.87109375,
      "loss": 0.7497,
      "rewards/accuracies": 0.4791666567325592,
      "rewards/chosen": 0.0066477167420089245,
      "rewards/margins": 0.033824387937784195,
      "rewards/rejected": -0.02717665769159794,
      "step": 10
    },
    {
      "epoch": 0.3791469194312796,
      "grad_norm": 151.6525661192009,
      "learning_rate": 4.885348141000122e-07,
      "logits/chosen": 121.6771240234375,
      "logits/rejected": 125.45533752441406,
      "logps/chosen": -372.70947265625,
      "logps/rejected": -423.9148864746094,
      "loss": 0.66,
      "rewards/accuracies": 0.5687500238418579,
      "rewards/chosen": -0.01487281359732151,
      "rewards/margins": 0.18130120635032654,
      "rewards/rejected": -0.1961739957332611,
      "step": 20
    },
    {
      "epoch": 0.5687203791469194,
      "grad_norm": 127.760794507037,
      "learning_rate": 4.5025027361734613e-07,
      "logits/chosen": 140.84371948242188,
      "logits/rejected": 134.668701171875,
      "logps/chosen": -408.40216064453125,
      "logps/rejected": -448.96710205078125,
      "loss": 0.6211,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.7968470454216003,
      "rewards/margins": 0.6513954997062683,
      "rewards/rejected": -1.4482426643371582,
      "step": 30
    },
    {
      "epoch": 0.7582938388625592,
      "grad_norm": 117.3200007673402,
      "learning_rate": 3.893311157806091e-07,
      "logits/chosen": 123.24369049072266,
      "logits/rejected": 112.3370361328125,
      "logps/chosen": -381.4453125,
      "logps/rejected": -402.53936767578125,
      "loss": 0.6035,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -1.3625555038452148,
      "rewards/margins": 0.8706506490707397,
      "rewards/rejected": -2.233206033706665,
      "step": 40
    },
    {
      "epoch": 0.9478672985781991,
      "grad_norm": 205.0988018258791,
      "learning_rate": 3.126631330646801e-07,
      "logits/chosen": 140.59500122070312,
      "logits/rejected": 144.99911499023438,
      "logps/chosen": -465.456298828125,
      "logps/rejected": -541.2003173828125,
      "loss": 0.5469,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -2.3156235218048096,
      "rewards/margins": 0.9280710220336914,
      "rewards/rejected": -3.243694305419922,
      "step": 50
    },
    {
      "epoch": 1.1374407582938388,
      "grad_norm": 103.63573519510389,
      "learning_rate": 2.2891223348923882e-07,
      "logits/chosen": 132.75270080566406,
      "logits/rejected": 136.41439819335938,
      "logps/chosen": -441.16741943359375,
      "logps/rejected": -510.4600524902344,
      "loss": 0.463,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -2.1580872535705566,
      "rewards/margins": 1.546640157699585,
      "rewards/rejected": -3.7047271728515625,
      "step": 60
    },
    {
      "epoch": 1.3270142180094786,
      "grad_norm": 88.9555531672405,
      "learning_rate": 1.4754491880085317e-07,
      "logits/chosen": 127.8622055053711,
      "logits/rejected": 129.76100158691406,
      "logps/chosen": -398.2987365722656,
      "logps/rejected": -478.58526611328125,
      "loss": 0.4087,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": -1.5577425956726074,
      "rewards/margins": 1.6958301067352295,
      "rewards/rejected": -3.253572940826416,
      "step": 70
    },
    {
      "epoch": 1.5165876777251186,
      "grad_norm": 91.77009057587127,
      "learning_rate": 7.775827023107834e-08,
      "logits/chosen": 115.83001708984375,
      "logits/rejected": 132.09730529785156,
      "logps/chosen": -392.9648742675781,
      "logps/rejected": -491.1802673339844,
      "loss": 0.3483,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -1.9656927585601807,
      "rewards/margins": 1.931434988975525,
      "rewards/rejected": -3.897127628326416,
      "step": 80
    },
    {
      "epoch": 1.7061611374407581,
      "grad_norm": 83.21809324441205,
      "learning_rate": 2.7440387297912122e-08,
      "logits/chosen": 115.92982482910156,
      "logits/rejected": 129.11094665527344,
      "logps/chosen": -424.06640625,
      "logps/rejected": -514.6354370117188,
      "loss": 0.3396,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": -1.8817684650421143,
      "rewards/margins": 1.9659159183502197,
      "rewards/rejected": -3.847684383392334,
      "step": 90
    },
    {
      "epoch": 1.8957345971563981,
      "grad_norm": 71.96649220747696,
      "learning_rate": 2.27878296044029e-09,
      "logits/chosen": 122.7137680053711,
      "logits/rejected": 122.6277084350586,
      "logps/chosen": -415.48455810546875,
      "logps/rejected": -488.2542419433594,
      "loss": 0.3434,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -1.6754734516143799,
      "rewards/margins": 1.8990375995635986,
      "rewards/rejected": -3.5745112895965576,
      "step": 100
    },
    {
      "epoch": 1.971563981042654,
      "step": 104,
      "total_flos": 0.0,
      "train_loss": 0.5053261174605443,
      "train_runtime": 2173.3813,
      "train_samples_per_second": 6.212,
      "train_steps_per_second": 0.048
    }
  ],
  "logging_steps": 10,
  "max_steps": 104,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}