MultiNLI-cleaned

--- dataset_info: features: - name: gold_label dtype: string - name: sentence1 dtype: string - name: sentence2 dtype: string splits: - name: train num_bytes: 75405059.0