{
    "epoch": 2.97,
    "eval_logits/chosen": -2.226133108139038,
    "eval_logits/rejected": -2.1600077152252197,
    "eval_logps/chosen": -264.3455505371094,
    "eval_logps/rejected": -222.62977600097656,
    "eval_loss": 0.6542457342147827,
    "eval_rewards/accuracies": 0.6765000224113464,
    "eval_rewards/chosen": 0.06409955769777298,
    "eval_rewards/margins": 0.08906211704015732,
    "eval_rewards/rejected": -0.024962568655610085,
    "eval_runtime": 2320.838,
    "eval_samples": 2000,
    "eval_samples_per_second": 0.862,
    "eval_steps_per_second": 0.215,
    "train_loss": 0.6692726473427482,
    "train_runtime": 42824.2925,
    "train_samples": 6196,
    "train_samples_per_second": 0.434,
    "train_steps_per_second": 0.007
}