{
    "epoch": 1.0,
    "eval_logits/chosen": 0.951361358165741,
    "eval_logits/rejected": 1.4587281942367554,
    "eval_logps/chosen": -426.3906555175781,
    "eval_logps/rejected": -491.477783203125,
    "eval_loss": 0.5107128024101257,
    "eval_rewards/accuracies": 0.77182537317276,
    "eval_rewards/chosen": -1.464455246925354,
    "eval_rewards/margins": 0.8910558819770813,
    "eval_rewards/rejected": -2.355511426925659,
    "eval_runtime": 244.9143,
    "eval_samples": 35044,
    "eval_samples_per_second": 8.166,
    "eval_steps_per_second": 0.257,
    "train_loss": 0.5487770005670517,
    "train_runtime": 16595.3532,
    "train_samples": 179264,
    "train_samples_per_second": 3.684,
    "train_steps_per_second": 0.058
}