{
    "epoch": 0.9995638901003053,
    "eval_logits/chosen": -2.924424171447754,
    "eval_logits/rejected": -2.9938888549804688,
    "eval_logps/chosen": -2.5334479808807373,
    "eval_logps/rejected": -3.1372740268707275,
    "eval_loss": 0.8931147456169128,
    "eval_rewards/accuracies": 0.7921686768531799,
    "eval_rewards/chosen": -0.36102983355522156,
    "eval_rewards/margins": 0.22724701464176178,
    "eval_rewards/rejected": -0.5882768034934998,
    "eval_runtime": 114.5732,
    "eval_samples": 2643,
    "eval_samples_per_second": 23.068,
    "eval_steps_per_second": 0.724,
    "total_flos": 0.0,
    "train_loss": 1.0578667073141634,
    "train_runtime": 8389.7514,
    "train_samples": 73350,
    "train_samples_per_second": 8.743,
    "train_steps_per_second": 0.068
}