{
    "epoch": 2.998693948628646,
    "eval_logits/chosen": 0.182608962059021,
    "eval_logits/rejected": 0.23953425884246826,
    "eval_logps/chosen": -338.5595703125,
    "eval_logps/rejected": -388.22418212890625,
    "eval_loss": 0.5813568234443665,
    "eval_rewards/accuracies": 0.7848101258277893,
    "eval_rewards/chosen": 1.7432384490966797,
    "eval_rewards/margins": 5.916741847991943,
    "eval_rewards/rejected": -4.173503398895264,
    "eval_runtime": 69.9912,
    "eval_samples": 2500,
    "eval_samples_per_second": 35.719,
    "eval_steps_per_second": 1.129,
    "total_flos": 0.0,
    "train_loss": 0.5855818307081304,
    "train_runtime": 16735.6732,
    "train_samples": 73493,
    "train_samples_per_second": 13.174,
    "train_steps_per_second": 0.103
}