CaseHOLD
收藏资源简介:
CaseHOLD数据集由斯坦福大学创建,包含超过53,000个多选问题,旨在帮助律师识别相关案例裁决。数据集通过提取美国法律案例中的引用和裁决声明构建,每个问题提供一个案例背景和多个可能的裁决选项,其中一个是正确答案。创建过程涉及从法律引用规则中提取信息,确保数据集的法律意义和难度。该数据集主要应用于法律自然语言处理领域,特别是在法律裁决理解和分析中,以提高法律专业人士的工作效率和准确性。
CaseHOLD dataset was developed by Stanford University, which contains over 53,000 multiple-choice questions designed to assist legal professionals in identifying relevant case holdings. The dataset is constructed by extracting citations and holding statements from U.S. legal cases. Each question provides a case background and several potential ruling options, with exactly one correct answer. The creation process involves extracting information from legal citation rules to ensure the dataset's legal relevance and appropriate difficulty level. This dataset is primarily applied in the field of legal natural language processing, particularly in legal ruling understanding and analysis, to improve the work efficiency and accuracy of legal professionals.

- 1When Does Pretraining Help? Assessing Self-Supervised Learning for Law and the CaseHOLD Dataset斯坦福大学 · 2021年



