Dataset Overview

Dataset Information

• Features:
  • book: name or code of the legal book (e.g., "Civil Code", "Penal Code")
  • document: full text of the legal document
  • timestamp: timestamp of enactment or last update
  • id: identifier of each document
  • hash: SHA‑256 hash of the document for verification purposes
• Splits:
  • fr: contains 153,005 samples, total size 151,400,300 bytes
• Download Size: 64,396,801 bytes
• Dataset Size: 151,400,300 bytes
• Configuration:
  • default: loads the fr split using the data/fr-* path
• License: cc‑by‑4.0
• Task Types:
  • Question Answering
  • Text Generation
  • Table Question Answering
• Language: French
• Tags:
  • Law
  • Law
  • Finance
  • Taxation
  • δεξιά
  • recht
  • derecho
• Name: The Laws, centralizing legal texts for better use

Objectives

• Centralize legal texts worldwide in a common format to facilitate:
  1. Comparative legal research
  2. Development of multilingual legal AI models
  3. Cross‑jurisdictional legal research
  4. Improvement of legal tech tools

Dataset Structure

• book: name or code of the legal book
• document: full text of the legal document
• timestamp: timestamp of enactment or last update
• id: identifier of each document
• hash: SHA‑256 hash of the document for verification purposes

Country Splits

• The dataset is organized by country using ISO 3166‑1 alpha‑2 codes:
  • France: fr
  • United States: us
  • United Kingdom: gb
  • Germany: de
  • Japan: jp
  • Brazil: br

Ethical Considerations

• Privacy: Ensure all personal information is properly anonymized.
• Bias: Be aware of potential biases in source materials and in the selection of laws.
• Timeliness: Laws evolve; always verify that the version you use is up‑to‑date.
• Jurisdiction: Legal interpretation may vary across jurisdictions. AI models trained on this data should not replace professional legal advice.

Citation

• If you use this dataset in your research, please cite the following BibTeX entry:

@misc{HFforLegal2024,
  author = {Louis Brulé Naudet},
  title = {The Laws, centralizing legal texts for better use},
  year = {2024},
  howpublished = {url{https://huggingface.co/datasets/HFforLegal/laws}}
}