all_results.json

{
    "epoch": 0.99,
    "eval_logits/chosen": -2.6191320419311523,
    "eval_logits/rejected": -2.66033935546875,
    "eval_logps/chosen": -198.21322631835938,
    "eval_logps/rejected": -216.69395446777344,
    "eval_loss": 3718.415283203125,
    "eval_rewards/accuracies": 0.5239361524581909,
    "eval_rewards/chosen": -0.424562007188797,
    "eval_rewards/margins": 0.069640152156353,
    "eval_rewards/rejected": -0.494202196598053,
    "eval_rewards/safe_rewards": -0.43978598713874817,
    "eval_rewards/unsafe_rewards": -0.4005235433578491,
    "eval_runtime": 140.0797,
    "eval_samples": 1487,
    "eval_samples_per_second": 10.615,
    "eval_steps_per_second": 0.336,
    "train_loss": 5218.799974524457,
    "train_runtime": 254.7215,
    "train_samples": 1487,
    "train_samples_per_second": 5.838,
    "train_steps_per_second": 0.181
}