Contract Understanding Atticus Dataset (CUAD)

Contract Understanding Atticus Dataset (CUAD) 是一个由The Atticus Project和加州大学伯克利分校合作创建的大型法律合同审查数据集。该数据集包含超过500份合同，总计超过13,000条由法律专家精心标注的重要条款，涵盖41种不同的标签类别。CUAD的创建过程涉及法律学生和律师的广泛培训，确保标注的质量和一致性。该数据集主要用于研究自然语言处理模型在高度专业化领域，如法律合同审查中的应用，旨在通过自动化技术减轻法律专业人士的工作负担，提高合同审查的效率。

Contract Understanding Atticus Dataset (CUAD) is a large-scale legal contract review dataset co-created by The Atticus Project and the University of California, Berkeley. This dataset contains over 500 contracts, totaling more than 13,000 critical clauses meticulously annotated by legal experts, covering 41 distinct label categories. The development of CUAD involved extensive training for law students and lawyers to ensure the quality and consistency of annotations. This dataset is primarily used for researching the application of natural language processing models in highly specialized domains such as legal contract review, aiming to alleviate the workload of legal professionals via automated technologies and improve the efficiency of contract review.