---
license: apache-2.0
language:
  - en
tags:
  - legal
  - legal-ai
  - reasoning
  - reinforcement-learning
  - large-language-models
datasets:
  - Xubqpanda/LegalDelta
---

# LegalΔ

This repository contains the model checkpoints for **LegalΔ: Enhancing Legal Reasoning in LLMs via Reinforcement Learning with Chain-of-Thought Guided Information Gain**.

Paper: https://arxiv.org/abs/2508.12281

Dataset: https://huggingface.co/datasets/Xubqpanda/LegalDelta

GitHub: https://github.com/NEUIR/LegalDelta

## Citation

```bibtex
@article{DBLP:journals/corr/abs-2508-12281,
  author       = {Xin Dai and
                  Buqiang Xu and
                  Zhenghao Liu and
                  Yukun Yan and
                  Huiyuan Xie and
                  Xiaoyuan Yi and
                  Shuo Wang and
                  Ge Yu},
  title        = {Legal{\(\Delta\)}: Enhancing Legal Reasoning in LLMs via Reinforcement
                  Learning with Chain-of-Thought Guided Information Gain},
  journal      = {CoRR},
  volume       = {abs/2508.12281},
  year         = {2025},
  url          = {https://doi.org/10.48550/arXiv.2508.12281},
  doi          = {10.48550/ARXIV.2508.12281},
  eprinttype   = {arXiv},
  eprint       = {2508.12281},
  timestamp    = {Tue, 16 Sep 2025 21:07:06 +0200},
  biburl       = {https://dblp.org/rec/journals/corr/abs-2508-12281.bib},
  bibsource    = {dblp computer science bibliography, https://dblp.org}
}