LH2-data-labs/indian-legal-records

该数据集是印度法律记录和判决语料库，提供了印度司法系统的结构化、索引化和部分标记的法律记录，规模之大无公开等效数据。覆盖印度最高法院、25个高等法院（包括所有分庭和管辖权）、600多个地区和下级法院，以及专门法庭如国家公司法法庭（NCLT）、中央行政法庭（CAT）和消费者纠纷解决论坛。数据集独特之处在于其AI增强层：每个法院命令都配有预先计算的简明摘要、提取的关键法律点、结果分类、具体救济措施和引用的法律条款。数据集设计用于直接输入AI训练管道，用于基础模型预训练、法律推理模型微调、检索增强生成（RAG）管道开发和评估基准构建。

This corpus provides structured, indexed, and partially labelled legal records from the Indian judicial system at a scale that has no public equivalent. It covers the Supreme Court of India, 25 High Courts (with all benches and jurisdictions), 600+ District and Subordinate Courts, and specialised tribunals including the National Company Law Tribunal (NCLT), Central Administrative Tribunal (CAT), and Consumer Dispute Redressal Forums. What distinguishes this corpus is the AI enrichment layer: every court order is paired with pre-computed plain-language summaries, extracted key legal points, outcome classification, specific relief granted, and cited legal provisions. The corpus is designed for direct ingestion into AI training pipelines for foundation model pre-training, legal reasoning model fine-tuning, retrieval-augmented generation (RAG) pipeline development, and evaluation benchmark construction.