korean-law-articles
收藏资源简介:
Korean Law Articles 数据集是一个包含韩国现行法律全文的全面数据集,共收集了5,583条法律(总数为5,584条,其中1条因来源服务器未提供正文而缺失)。该数据集通过韩国法制处国家法令信息OPEN API(`lawService.do`)收集,适用于文本生成、文本分类、问答、摘要和特征提取等多种自然语言处理任务。数据集包含丰富的字段信息,如法律ID、法令名称、法令类型、主管部门、公布日期、施行日期、条文内容等。每条法律记录还包括条文列表(编号、标题、内容、施行日期)和全文文本(所有条文合并)。数据集规模适中,平均每条法律包含约8个条文,全文平均长度约2,200字。数据集采用CC0-1.0许可,要求使用时注明来源。该数据集可用于韩国法律领域的大型语言模型预训练/微调、法令检索/嵌入基准测试、法律问答系统评估、法律NER/摘要标注种子数据以及主管部门分析等场景。数据集每6小时自动更新一次,确保包含最新的法律修订内容。
The Korean Law Articles Dataset is a comprehensive dataset containing the full texts of currently effective South Korean laws. A total of 5,583 legal documents were collected, with the original total being 5,584, and 1 law missing because its full text was not provided by the source server. This dataset was collected via the National Law Information OPEN API (`lawService.do`) provided by the Ministry of Justice of the Republic of Korea. It supports a wide range of natural language processing (NLP) tasks, including text generation, text classification, question answering, text summarization, and feature extraction. The dataset includes rich field information such as law ID, decree name, decree type, competent authority, promulgation date, enforcement date, and article content. Each law record also includes an article list (with article number, title, content, and enforcement date) and the full merged text of all articles. The dataset is of moderate scale, with an average of approximately 8 articles per law and an average full text length of about 2,200 words. It is licensed under CC0-1.0, and users are required to indicate the source when using the dataset. Potential use cases of this dataset include pre-training or fine-tuning large language models (LLMs) for the South Korean legal domain, legal document retrieval and embedding benchmark testing, evaluation of legal question answering systems, seed data for legal NER and summarization annotation, and competent authority analysis. The dataset is automatically updated every 6 hours to ensure it contains the latest legal revisions.




