{"url":"/dataset/xsum","name":"XSum","full_name":null,"description_markdown":"The Extreme Summarization (**XSum**) dataset is a dataset for evaluation of abstractive single-document summarization systems. The goal is to create a short, one-sentence new summary answering the question “What is the article about?”. The dataset consists of 226,711 news articles accompanied with a one-sentence summary. The articles are collected from BBC articles (2010 to 2017) and cover a wide variety of domains (e.g., News, Politics, Sports, Weather, Business, Technology, Science, Health, Family, Education, Entertainment and Arts). The official random split contains 204,045 (90%), 11,332 (5%) and 11,334 (5) documents in training, validation and test sets, respectively.\n\nSource: [https://arxiv.org/pdf/1808.08745.pdf](https://arxiv.org/pdf/1808.08745.pdf)\nImage Source: [https://arxiv.org/pdf/1808.08745.pdf](https://arxiv.org/pdf/1808.08745.pdf)","description_withheld":null,"homepage":"https://github.com/EdinburghNLP/XSum/tree/master/XSum-Dataset","introduced_date":"2018-01-01","introduced_date_note":null,"introduced_by":{"paper":"/paper/dont-give-me-the-details-just-the-summary","title":"Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarization","first_author":"Shashi Narayan","url":null},"license":null,"modalities":[{"name":"Texts","url":"/datasets/modality/texts"}],"tasks":[{"name":"Text Summarization","url":"/task/text-summarization","datasets_with_task":"/datasets/task/text-summarization"},{"name":"Abstractive Text Summarization","url":"/task/abstractive-text-summarization","datasets_with_task":"/datasets/task/abstractive-text-summarization"},{"name":"Sequence-to-sequence Language Modeling","url":"/task/sequence-to-sequence-language-modeling","datasets_with_task":"/datasets/task/sequence-to-sequence-language-modeling"},{"name":"Summarization","url":"/task/summarization","datasets_with_task":"/datasets/task/summarization"},{"name":"Extreme Summarization","url":"/task/extreme-summarization","datasets_with_task":"/datasets/task/extreme-summarization"}],"languages":[{"name":"English","url":"/datasets/language/english"}],"variants":["XSum","X-Sum"],"data_loaders":[{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/talgatzh/xfinetuning1","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/xsum","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/talgatzh/xsum-kk3","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/shalinik/xsum","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/lewtun/autoevaluate__xsum","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/polinaeterna/xsum","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/huggingface/datasets","url":"https://huggingface.co/datasets/EdinburghNLP/xsum","frameworks":["tf","pytorch","jax"]},{"repo":"https://github.com/tensorflow/datasets","url":"https://www.tensorflow.org/datasets/catalog/xsum","frameworks":["tf","jax"]},{"repo":"https://github.com/EdinburghNLP/XSum","url":"https://github.com/EdinburghNLP/XSum","frameworks":["pytorch"]}],"num_papers_in_archive":32,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/text-summarization-on-x-sum","task":"Text Summarization","dataset_variant":"X-Sum","rows":18,"metrics":["ROUGE-1","ROUGE-2","ROUGE-3","ROUGE-L"],"first_row_in_archive_order":{"model":"Selfmem","paper":"/paper/lift-yourself-up-retrieval-augmented-text","metrics":{"ROUGE-1":"50.30","ROUGE-2":"26.70","ROUGE-3":"41.60"},"code_links":[{"title":"hannibal046/selfmemory","url":"https://github.com/hannibal046/selfmemory"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/extreme-summarization-on-xsum","task":"Extreme Summarization","dataset_variant":"XSum","rows":1,"metrics":["METEOR"],"first_row_in_archive_order":{"model":"PEGASUS","paper":"/paper/the-gem-benchmark-natural-language-generation","metrics":{"METEOR":"0.216"},"code_links":[]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/text-summarization-on-xsum-2","task":"Text Summarization","dataset_variant":"XSum","rows":1,"metrics":["ROUGE-1"],"first_row_in_archive_order":{"model":"SRformer-BART","paper":"/paper/segmented-recurrent-transformer-an-efficient","metrics":{"ROUGE-1":"39.02"},"code_links":[{"title":"yinghanlong/SRtransformer","url":"https://github.com/yinghanlong/SRtransformer"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/abstractive-text-summarization-on-xsum","task":"Abstractive Text Summarization","dataset_variant":"XSum","rows":0,"metrics":["Test ROGUE-1","Test ROGUE-2","Test ROGUE-L","Test ROGUE-Lsum","Validation ROGUE-1","Validation ROGUE-2","Validation ROGUE-L","Validation ROGUE-Lsum"],"first_row_in_archive_order":null,"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/summarization-on-xsum","task":"Summarization","dataset_variant":"XSum","rows":0,"metrics":["ROUGE-1","ROUGE-2","ROUGE-L","ROUGE-LSUM","gen_len","loss"],"first_row_in_archive_order":null,"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/segmented-recurrent-transformer-an-efficient","title":"Segmented Recurrent Transformer: An Efficient Sequence-to-Sequence Model","date":"2023-05-24","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/palm-2-technical-report-1","title":"PaLM 2 Technical Report","date":"2023-05-17","rows_on_this_dataset":3,"code_links":1,"syntology":null},{"paper":"/paper/lift-yourself-up-retrieval-augmented-text","title":"Lift Yourself Up: Retrieval-augmented Text Generation with Self Memory","date":"2023-05-03","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":1,"samples_ran":1,"samples_unverified":0,"pointer_only_for_licence":1,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/brio-bringing-order-to-abstractive","title":"BRIO: Bringing Order to Abstractive Summarization","date":"2022-03-31","rows_on_this_dataset":1,"code_links":3,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":10,"samples_ran":0,"samples_unverified":10,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/summareranker-a-multi-task-mixture-of-experts-1","title":"SummaReranker: A Multi-Task Mixture-of-Experts Re-ranking Framework for Abstractive Summarization","date":"2022-03-13","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":3,"samples_ran":2,"samples_unverified":1,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/simcls-a-simple-framework-for-contrastive","title":"SimCLS: A Simple Framework for Contrastive Learning of Abstractive Summarization","date":"2021-06-03","rows_on_this_dataset":1,"code_links":2,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":6,"samples_ran":4,"samples_unverified":2,"pointer_only_for_licence":1,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/hierarchical-learning-for-generation-with","title":"Hierarchical Learning for Generation with Long Source Sequences","date":"2021-04-15","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/the-gem-benchmark-natural-language-generation","title":"The GEM Benchmark: Natural Language Generation, its Evaluation and Metrics","date":"2021-02-02","rows_on_this_dataset":1,"code_links":0,"syntology":null},{"paper":"/paper/pegasus-pre-training-with-extracted-gap","title":"PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization","date":"2019-12-18","rows_on_this_dataset":1,"code_links":19,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":19,"samples_ran":1,"samples_unverified":18,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/bart-denoising-sequence-to-sequence-pre","title":"BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension","date":"2019-10-29","rows_on_this_dataset":1,"code_links":47,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":53,"samples_ran":22,"samples_unverified":31,"pointer_only_for_licence":7,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/text-summarization-with-pretrained-encoders","title":"Text Summarization with Pretrained Encoders","date":"2019-08-22","rows_on_this_dataset":1,"code_links":19,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":21,"samples_ran":7,"samples_unverified":14,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/dont-give-me-the-details-just-the-summary","title":"Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarization","date":"2018-08-27","rows_on_this_dataset":7,"code_links":3,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":4,"samples_ran":0,"samples_unverified":4,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":8,"samples_harvested":117,"samples_ran":37,"samples_unverified":80,"pointer_only_for_licence":9,"papers_with_no_sample_that_ran":2,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}