task_categories: - summarization pretty_name: PubMed 10k size_categories: - 1K<n<10K
First 10k rows of the scientific_papers["pubmed"] dataset. 10:1:1 split.