{
    "epoch": 2.986666666666667,
    "eval_logits/chosen": -0.3865179717540741,
    "eval_logits/rejected": -0.45662835240364075,
    "eval_logps/chosen": -0.15898314118385315,
    "eval_logps/rejected": -0.6109744310379028,
    "eval_loss": 0.20052169263362885,
    "eval_odds_ratio_loss": 1.8274307250976562,
    "eval_rewards/accuracies": 0.8299999833106995,
    "eval_rewards/chosen": -0.015898313373327255,
    "eval_rewards/margins": 0.045199137181043625,
    "eval_rewards/rejected": -0.06109744682908058,
    "eval_runtime": 3.8864,
    "eval_samples_per_second": 25.73,
    "eval_sft_loss": 0.01777861826121807,
    "eval_steps_per_second": 12.865,
    "total_flos": 2.94774302048256e+16,
    "train_loss": 0.703327341448693,
    "train_runtime": 380.828,
    "train_samples_per_second": 7.09,
    "train_steps_per_second": 0.441
}