This repository is a curated list of resources for Legal Natural Language Processing (LegalNLP), targeting researchers and practitioners in law, computer science, and AI. It provides a comprehensive overview of datasets, benchmarks, models, and academic literature, facilitating the development and application of NLP techniques to legal texts.

How It Works

The resource is organized by NLP task within the legal domain, including Legal Judgement Prediction, Legal Text Classification, Legal Information Retrieval, Legal Question Answering, Legal Textual Entailment, Legal Text Summarization, and Legal Language Modeling. For each task, it lists relevant datasets with links to access them, their domain, language, and size. It also highlights key benchmarks and pre-trained models tailored for legal NLP.

Quick Start & Requirements

This is a curated list, not a software package. Accessing datasets and models requires following individual links and adhering to their respective requirements. Many datasets are available via Hugging Face (🤗) or direct download (💾).

Highlighted Details

• Comprehensive coverage of 7 distinct LegalNLP tasks.
• Includes datasets in multiple languages (English, German, French, Italian, Spanish, Chinese, Hebrew, Japanese).
• Features prominent benchmarks like FairLex and LexGLUE for evaluating LegalNLP models.
• Lists various pre-trained legal language models, including Legal-BERT, JuriBERT, and LEGAL-GPT variants.

Maintenance & Community

The list is curated by maastrichtlawtech. It references academic papers, conferences (e.g., ICAIL, JURIX), and workshops (e.g., NLLP, XAILA), indicating a connection to the active research community in AI and Law.

Licensing & Compatibility

The repository itself is not software and does not have a license. Individual datasets and models listed will have their own licenses, which must be checked for compatibility with commercial or closed-source use.

Limitations & Caveats

This is a static list of resources; it does not provide direct access to the data or models, nor does it offer tools for integration. Users must navigate to each resource's source for download and usage instructions.