{
    "epoch": 1.0,
    "eval_logits/chosen": -1.1779704093933105,
    "eval_logits/rejected": -1.2511389255523682,
    "eval_logps/chosen": -335.2599792480469,
    "eval_logps/rejected": -317.35614013671875,
    "eval_loss": 0.6636306643486023,
    "eval_rewards/accuracies": 0.6746031641960144,
    "eval_rewards/chosen": 0.04071044921875,
    "eval_rewards/margins": 0.0732826218008995,
    "eval_rewards/rejected": -0.032572176307439804,
    "eval_runtime": 91.7551,
    "eval_samples": 2000,
    "eval_samples_per_second": 21.797,
    "eval_steps_per_second": 0.687,
    "total_flos": 0.0,
    "train_loss": 0.6732059223382543,
    "train_runtime": 6951.1203,
    "train_samples": 61134,
    "train_samples_per_second": 8.795,
    "train_steps_per_second": 0.034
}