{ "epoch": 2.999098751126561, "eval_kl/chosen": 0.000217721943045035, "eval_kl/rejected": 0.0005314685986377299, "eval_logits/chosen": -1.421875, "eval_logits/rejected": -1.390625, "eval_logps/chosen": -322.0, "eval_logps/rejected": -298.0, "eval_loss": 0.6974413990974426, "eval_nll_loss": 1.046875, "eval_ref_probs/chosen": 0.0038271306548267603, "eval_ref_probs/rejected": 0.0034569373819977045, "eval_rejected_term/max": 0.0005314685986377299, "eval_rejected_term/min": 0.0005314685986377299, "eval_rewards/accuracies": 0.6000000238418579, "eval_rewards/chosen": -0.01556396484375, "eval_rewards/margins": 0.13671875, "eval_rewards/rejected": -0.1513671875, "eval_runtime": 25.3373, "eval_samples_per_second": 39.468, "eval_steps_per_second": 9.867 }