labels mask

--- dataset_info: features: - name: step dtype: string - name: label dtype: string splits: - name: train num_bytes: 4985662 num_examples: 23330 - name: test num_bytes: