all_results.json

{
    "epoch": 3.0,
    "eval_logits/chosen": -2.2820301055908203,
    "eval_logits/rejected": -2.246671438217163,
    "eval_logps/chosen": -244.1195068359375,
    "eval_logps/rejected": -238.75563049316406,
    "eval_loss": 0.5447676181793213,
    "eval_rewards/accuracies": 0.796875,
    "eval_rewards/chosen": 0.42937949299812317,
    "eval_rewards/margins": 2.895768404006958,
    "eval_rewards/rejected": -2.4663889408111572,
    "eval_runtime": 203.3244,
    "eval_samples_per_second": 14.951,
    "eval_steps_per_second": 0.236,
    "total_flos": 3785055088410624.0,
    "train_loss": 0.32465138984989156,
    "train_runtime": 12160.5022,
    "train_samples_per_second": 6.748,
    "train_steps_per_second": 0.026
}