Tim-Saijun/grpo_n4_bz512_120

--- dataset_info: features: - name: problem dtype: string - name: answers sequence: string - name: gt dtype: string - name: data_source dtype: string splits: - name: train num_bytes: 43540046 num_examples: 1517 download_size: 13947257 dataset_size: 43540046 configs: - config_name: default data_files: - split: train path: data/train-* ---