{
    "epoch": 1.0,
    "eval_logits/chosen": -0.21594780683517456,
    "eval_logits/rejected": 2.576901435852051,
    "eval_logps/chosen": -609.2966918945312,
    "eval_logps/rejected": -7672.10595703125,
    "eval_loss": 7.102515519363806e-05,
    "eval_rewards/accuracies": 1.0,
    "eval_rewards/chosen": -2.423149824142456,
    "eval_rewards/margins": 67.89002990722656,
    "eval_rewards/rejected": -70.31317901611328,
    "eval_runtime": 191.0845,
    "eval_samples": 3905,
    "eval_samples_per_second": 20.436,
    "eval_steps_per_second": 0.324
}