XL-Sum

该数据集名为XL-Sum，是一个大规模的多语言摘要数据集，包含了45种语言的标注文档-摘要对集合。该数据集用于评估多种语言摘要的忠实度，并支持开展检测虚构内容的实验。其规模之大，涵盖了45种语言的数据，所涉及的任务是抽象式摘要。This dataset, named XL-Sum, is a large-scale multilingual summarization dataset containing a collection of annotated document-summary pairs across 45 languages. It is designed to evaluate the faithfulness of multilingual summarization models and supports experiments for detecting fabricated content. Spanning 45 languages, this dataset targets the task of abstractive summarization.