Panoramic Cross-Research between Large Language Models and Law: A Paper Repository from Legal Reasoning to Intelligent Legal Assistants

The LLM-and-Law repository is maintained by Jeryi-Sun, sourced from GitHub, released on May 30, 2026, and updated daily. This repository systematically organizes papers on the applications of large language models in the legal field, covering seven dimensions: legal task applications, legal reasoning models, legal agents, legal issue research, data resources, legal domain models, and evaluation benchmarks. It provides a comprehensive resource index for researchers, practitioners, and policymakers.

The legal industry, with its dense rules and complex texts, is an ideal testbed for LLM applications, but it has unique requirements such as the rigor of legal texts, the specificity of reasoning, and the social impact of decisions. In recent years, relevant research results have been scattered, so the LLM-and-Law project emerged to systematically track and summarize related papers.

The repository is divided into seven categories: 1. Legal Task Applications (judgment prediction, document summarization, etc.); 2. Legal Reasoning Models (prompt design, precedent enhancement, etc.); 3. Legal Agents (execution of complex processes); 4. Legal Issues (copyright, liability attribution, etc.); 5. Data Resources (judgment databases, regulatory corpora, etc.); 6. Legal Domain Models (LawGPT, Legal-BERT, etc.); 7. Evaluation Benchmarks (LexGLUE, Legal-NLI, etc.).

The project is updated daily. The maintainer regularly scans academic platforms to screen new papers, using strict criteria: only papers involving both "legal tasks" and "LLM semantics" are included, ensuring the quality and relevance of resources.

For AI researchers: It provides legal application scenarios and problem definitions; For legal practitioners: It showcases the latest technological progress and helps judge tool usability; For policymakers: It provides a research perspective on AI legal issues and assists in formulating regulatory frameworks.

Recommended path for researchers: 1. Start with application papers; 2. Dive into reasoning model research; 3. Focus on evaluation benchmarks; 4. Track legal issue research. In the future, with the advancement of LLM technology and the digital transformation of law, more results will emerge in the cross field, and the project will continue to provide value.

The integration of LLMs and law is full of opportunities and challenges. The LLM-and-Law project builds knowledge infrastructure for this field and is worth collecting and following by AI researchers, legal practitioners, and legal tech observers.