OpenMLRL/arXiv_abstract

arXiv Abstract This dataset is based on the arXiv scientific papers and is used for the text expansion task. (Download raw data here). I processed the raw data for the article expansion task with extract_arXiv_abstract.py. The processed dataset only contains the article ID and abstract fields, and the abstract length should be 100-300 tokens. The JSON objects are in the following format: { 'article_id': str, 'abstract_text': List[str], 'token_count': int }