CrossRE

该数据集名为CrossRE，是一个跨越句子级别的跨领域关系提取数据集。它包含了手工标注的17种关系类型，涵盖了新闻、政治、自然科学、音乐、文学和人工智能等六个不同文本领域，并允许进行多标签注释。此外，该数据集的注释还丰富了元数据信息，如解释和语法歧义。其规模在各个文本领域间具有多样性，关系数量也各不相同。该数据集的主要任务是关系提取。

This dataset, named CrossRE, is a sentence-level cross-domain relation extraction dataset. It contains 17 manually annotated relation types, covering six distinct text domains including news, politics, natural sciences, music, literature, and artificial intelligence, and supports multi-label annotation. Furthermore, its annotations are enriched with metadata such as explanatory notes and grammatical ambiguities. The dataset exhibits diversity in scale across various text domains, with the number of relations varying per domain. The primary task supported by this dataset is relation extraction.