eval_results.json

{
    "epoch": 3.99,
    "eval_logits/chosen": 0.81451815366745,
    "eval_logits/rejected": 0.8721050024032593,
    "eval_logps/chosen": -256.90045166015625,
    "eval_logps/rejected": -233.8546905517578,
    "eval_loss": 0.00017304150969721377,
    "eval_rewards/accuracies": 0.5130000114440918,
    "eval_rewards/chosen": -0.0028624406550079584,
    "eval_rewards/margins": 0.0003016398404724896,
    "eval_rewards/rejected": -0.003164080437272787,
    "eval_runtime": 412.2426,
    "eval_samples": 2000,
    "eval_samples_per_second": 4.852,
    "eval_steps_per_second": 1.213
}