{
    "epoch": 1.0,
    "eval_logits/chosen": 0.4467492401599884,
    "eval_logits/rejected": 1.781506061553955,
    "eval_logps/chosen": -380.6148681640625,
    "eval_logps/rejected": -458.5203552246094,
    "eval_loss": 0.49494317173957825,
    "eval_rewards/accuracies": 0.7890625,
    "eval_rewards/chosen": -1.0109279155731201,
    "eval_rewards/margins": 0.9840108156204224,
    "eval_rewards/rejected": -1.9949387311935425,
    "eval_runtime": 90.1258,
    "eval_samples": 2000,
    "eval_samples_per_second": 22.191,
    "eval_steps_per_second": 0.355,
    "train_loss": 0.08084109238500875,
    "train_runtime": 794.8705,
    "train_samples": 61155,
    "train_samples_per_second": 76.937,
    "train_steps_per_second": 0.601
}