new_causal_data_preprocessed/dataset_dict.json

{"splits": ["train", "validation", "test"]}