key,value root_checkpoint_step,90 root_checkpoint_kind,best_validation root_actor_dir,checkpoints/datasets/sciknoweval/biology/qwen3gen-biology-RLSD_TR-Qwen-Qwen3-4B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/_actor_archive/global_step_90/actor last_checkpoint_step,100 last_actor_dir,checkpoints/datasets/sciknoweval/biology/qwen3gen-biology-RLSD_TR-Qwen-Qwen3-4B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_100/actor best_validation_mean16,0.485000000000 final_validation_mean16,0.428750000000