mbti clean with categories

--- dataset_info: features: - name: label dtype: int64 - name: text dtype: string splits: - name: train num_bytes: 51651122 num_examples: 327828 - name: test num_bytes: