Jeryi-Sun/LLM-and-Law

What it solves

Keeping up with the fast-growing body of research on large language models (LLMs) applied to law is hard. Papers appear daily across many subtopics, from legal reasoning and retrieval to regulation and evaluation. This project solves that by maintaining a curated, categorized, and carefully filtered list of relevant papers and resources, with clear notes on what each paper does and its limitations.

How it works

The repository is a hand-maintained list of papers, organized into seven categories: Applications, Legal Reasoning Models, Legal Agent, Legal Problems, Data Resources, Law LLMs, and Evaluation. It is updated daily from arXiv listings (specifically the cs.CL category). Each daily update scans a batch of papers, applies a strict inclusion rule (papers must involve an explicit legal task and use a language model, agent, RAG, dataset, or benchmark), and adds title-deduplicated entries with brief annotations. The annotations often clarify that a paper's reported metrics do not guarantee legal correctness, and they exclude generic compliance or non-legal uses of legal terms.

Who it’s for

Researchers, practitioners, and students who work at the intersection of natural language processing and law and want a dependable, up-to-date bibliography. It is also useful for anyone tracking trends in legal LLM applications, evaluation, and data resources.

Highlights

• Daily updates with explicit scan counts and category increments.
• Strict inclusion criteria that keep out generic or tangential papers.
• Each entry links to the arXiv paper and includes a short, critical note.
• Covers a wide range of legal tasks across many languages and jurisdictions.