PMadesh006/filtered-cuad

过滤版合同理解阿提库斯数据集（filtered_cuad）是CUAD数据集的一个过滤版本，专注于商业法律合同中的关键条款识别。该数据集包含510份商业法律合同中的13,000多个标签，经过人工标注以识别41类重要条款。过滤版排除了2002年之前的合同和非B2B合同，并将类别缩减至12个最关键的类别。数据集旨在支持法律合同审查的NLP研究和开发，减少合同审查的社会成本，并提高小企业和个人获取法律支持的可及性。数据集为英文单语，包含训练集和测试集，分别有5,442和936个样本。

The filtered Contract Understanding Atticus Dataset (filtered_cuad) is a refined version of the CUAD dataset, focusing on identifying key clauses in commercial legal contracts. It comprises over 13,000 labels across 510 commercial legal contracts, manually annotated to identify 41 categories of important clauses. The filtered version excludes contracts dated before 2002 and non-B2B contracts, narrowing down to 12 most crucial categories. The dataset supports NLP research and development in legal contract review, aiming to reduce the societal costs of contract review and enhance access to legal support for small businesses and individuals. It is monolingual (English) and includes train and test splits with 5,442 and 936 samples, respectively.