{
    "epoch": 1.0,
    "eval_logits/chosen": 3.0354480743408203,
    "eval_logits/rejected": 3.5330193042755127,
    "eval_logps/chosen": -362.5973815917969,
    "eval_logps/rejected": -460.4252014160156,
    "eval_loss": 0.5147865414619446,
    "eval_rewards/accuracies": 0.765625,
    "eval_rewards/chosen": -0.9764143824577332,
    "eval_rewards/margins": 0.9741156101226807,
    "eval_rewards/rejected": -1.950529932975769,
    "eval_runtime": 39.7272,
    "eval_samples": 2000,
    "eval_samples_per_second": 50.343,
    "eval_steps_per_second": 0.805,
    "train_loss": 0.5581105443723032,
    "train_runtime": 5172.5891,
    "train_samples": 61134,
    "train_samples_per_second": 11.819,
    "train_steps_per_second": 0.092
}