ECtHR

该数据集是针对长文档分类的一个流行数据集，包含了欧洲人权法院的案件，并且每个段落都附有理由标注。此外，每个案件都与据称被违反的一个或多个公约条款相关联。该数据集的规模包括11,000个案件，分为8,866个训练样本，973个验证样本以及986个测试样本，其任务是进行多标签分类。

This is a popular dataset for long document classification, which consists of cases from the European Court of Human Rights. Each paragraph in these cases is accompanied by a rationale annotation. Additionally, each case is linked to one or more Convention articles that are alleged to have been violated. The dataset includes 11,000 total cases, split into 8,866 training samples, 973 validation samples, and 986 test samples, with the task being multi-label classification.