🌟 Hiring 🌟 The THUNLP group at Tsinghua University has internship opportunities in legal AI. Master/PhD students in computer science or law are very welcome to reach out via email.

⚖️ [2026-04] Our paper LexRel: Benchmarking Legal Relation Extraction for Chinese Civil Cases has been accepted to ACL 2026.

🤖 [2025-11] Our paper Large Language Models' Complicit Responses to Illicit Instructions across Socio-Legal Contexts is out!

🔗 [2025-09] Our paper LexChain: Modeling Legal Reasoning Chains for Chinese Tort Case Analysis has been accepted to AAAI 2026.

About Me

I am a postdoc researcher in the Tsinghua University NLP (THUNLP) Group. My current research interests sit at the intersection of AI and social sciences, with a particular interest in legal AI and computational modelling of legal reasoning. I am especially interested in building and evaluating legal LLMs: how they represent legal knowledge, reason over facts and legal rules, and support high-stakes legal decision-making.

At Tsinghua, I work with Prof. Zhiyuan Liu and Dr. Yuxiao Ye on AI and Law. Before joining Tsinghua, I was a Research Associate at the Cambridge Faculty of Law, working with Prof. Felix Steffek on dispute resolution for the UK Employment Tribunal. I also worked at the Centre for Business Research, University of Cambridge, on Legal Systems and Artificial Intelligence.

I completed my PhD in Computer Science in the NLIP group at the University of Cambridge, supervised by Prof. Ann Copestake. My PhD thesis studied robust evaluation for natural language generation, including grammaticality, faithfulness, and diversity.

I previously served as Chair and Treasurer of Women@CL at the University of Cambridge, and as Chief Operating Officer of Downing Enterprise at Downing College Cambridge. I have also helped organise the Cambridge PhD Multidisciplinary Seminars since 2019. Outside the lab, I enjoy playing squash, hiking, musicals, and ballet.

Selected Publications

For the most up-to-date list, please also see my Google ScholarGoogle Scholar profile.

(* indicates equal contribution; # indicates corresponding author)

LexRel: Benchmarking Legal Relation Extraction for Chinese Civil Cases

[paper] [code]

Yida Cai, Ranjuexiao Hu, Huiyuan Xie#, Chenyang Li, Yun Liu, Yuxiao Ye, Zhenghao Liu, Weixing Shen, Zhiyuan Liu#

ACL 2026

LexChain: Modeling Legal Reasoning Chains for Chinese Tort Case Analysis

[paper] [code]

Huiyuan Xie*, Chenyang Li*, Huining Zhu, Chubin Zhang, Yuxiao Ye#, Zhenghao Liu, Zhiyuan Liu#

AAAI 2026

Large Language Models' Complicit Responses to Illicit Instructions across Socio-Legal Contexts

[paper] [code]

Xing Wang*, Huiyuan Xie*, Yiyan Wang, Chaojun Xiao, Huimin Chen#, Holli Sargeant, Felix Steffek, Jie Shao, Zhiyuan Liu, Maosong Sun

Technical Report, ArXiv 2025

Information Extraction from Employment Tribunal Judgments Using A Large Language Model

[paper]

Joana de Faria, Huiyuan Xie, Felix Steffek#

Artificial Intelligence and Law, 2025

The CLC-UKET Dataset: Benchmarking Case Outcome Prediction for the UK Employment Tribunal

[paper]

Huiyuan Xie, Felix Steffek, Joana de Faria, Christine Carter, Jonathan Rutherford

NLLP Workshop at EMNLP 2024

The Cambridge Law Corpus: A Corpus for Legal AI Research

[paper] [website]

Andreas Ostling, Holli Sargeant, Huiyuan Xie, Ludwig Bull, Alexander Terenin, Leif Jonsson, Mans Magnusson, Felix Steffek

NeurIPS 2023

TIAGE: A Benchmark for Topic-Shift Aware Dialog Modeling

[paper] [code]

Huiyuan Xie, Zhenghao Liu, Chenyan Xiong, Zhiyuan Liu, Ann Copestake

Findings of EMNLP 2021