{
    "epoch": 1.0,
    "eval_logits/chosen": -2.2434139251708984,
    "eval_logits/rejected": -2.2385354042053223,
    "eval_logps/chosen": -34.0224494934082,
    "eval_logps/rejected": -37.51002883911133,
    "eval_loss": 0.9980219602584839,
    "eval_rewards/accuracies": 0.5141196250915527,
    "eval_rewards/chosen": 0.0048410315066576,
    "eval_rewards/margins": 0.002204673830419779,
    "eval_rewards/rejected": 0.0026363590732216835,
    "eval_runtime": 145.7351,
    "eval_samples": 343,
    "eval_samples_per_second": 2.354,
    "eval_steps_per_second": 0.295
}