Statutory Interpretation Data Set

该数据集包含了来自美国法典不同法规条款的42个短语，以及从提及这些短语的法院判决中提取的26,959个句子。这些句子根据其对解释的有用性被分类，本工作特别关注被标记为高价值的句子。此外，该语料库允许分析GPT-4基于高价值句子提供的语境，生成法律术语准确且相关的解释的能力。该数据集规模涉及1,853个与42个短语相关的高价值句子，任务旨在评估由GPT-4生成的法律术语解释的准确性。

This dataset comprises 42 phrases drawn from diverse statutory provisions of the United States Code, alongside 26,959 sentences extracted from court opinions that cite these phrases. These sentences are classified according to their utility for legal interpretation, with this work focusing particularly on the sentences labeled as high-value. Additionally, this corpus facilitates the analysis of GPT-4's capacity to produce accurate and relevant explanations for legal terminology based on the context provided by these high-value sentences. The dataset consists of 1,853 high-value sentences associated with the 42 phrases, and the task is designed to assess the accuracy of legal term explanations generated by GPT-4.