About me

Hello! I’m Yuejin Xie (谢悦晋), an undergraduate student majoring in Electronic Engineering at Huazhong University of Science and Technology (HUST). I will join IIGROUP, Tsinghua Shenzhen International Graduate School as a Master’s student, advised by Prof. Yujiu Yang.

My research interests focus on LLM & Agent Safety, including tool-use safety, agent guardrails, and alignment.

You can find my work on GitHub and Google Scholar.

News

• 2026.06: PaSBench-Video is now available on arXiv.
• 2026.05: AgentDoG 1.5 is now available on arXiv.
• 2026.04: ATBench, ATBench-Claw and ATBench-Codex, and Rethinking Generalization in Reasoning SFT are now available on arXiv.
• 2026.03: Code2Math is now available on arXiv.
• 2026.02: Frontier AI Risk Management Framework in Practice v1.5 is now available on arXiv.
• 2026.01: AgentDoG is now available on arXiv.
• 2025.12: Started internship at Shanghai AI Lab.
• 2025.09: One paper accepted by NeurIPS 2025 D&B Track.
• 2025.09: One paper accepted by EMNLP 2025 Main Conference.

Publications

• PaSBench-Video: A Streaming Video Benchmark for Proactive Safety Warning
  Yusong Zhao*, Yuejin Xie*, Youliang Yuan, Junjie Hu, Jitian Guo, Yujiu Yang, Pinjia He
  arXiv 2026 (Co-first Author) | [Paper] [Dataset]

• ATBench: A Diverse and Realistic Agent Trajectory Benchmark for Safety Evaluation and Diagnosis
  Yu Li*, Haoyu Luo*, Yuejin Xie*, Yuqian Fu, Zhonghao Yang, Shuai Shao, Qihan Ren, Wanying Qu, Yanwei Fu, Yujiu Yang, Jing Shao, Xia Hu, Dongrui Liu
  arXiv 2026 (Co-first Author) | [Paper] [Dataset]

• Code2Math: Can Your Code Agent Effectively Evolve Math Problems Through Exploration?
  Dadi Guo*, Yuejin Xie*, Qingyu Liu, Jiayu Liu, Zhiyuan Fan, Qihan Ren, Shuai Shao, Tianyi Zhou, Dongrui Liu, Yi R. Fung
  arXiv 2026 (Co-first Author) | [Paper] [Code]

• ToolSafety: A Comprehensive Dataset for Enhancing Safety in LLM-Based Agent Tool Invocations
  Yuejin Xie, Youliang Yuan, Wenxuan Wang, Fan Mo, Jianmin Guo, Pinjia He
  EMNLP 2025 Main Conference (First Author) | [Paper] [Code] [Dataset]

• AgentDoG 1.5: A Lightweight and Scalable Alignment Framework for AI Agent Safety and Security
  Dongrui Liu, Yu Li, Zhonghao Yang, Peng Wang, Guanxu Chen, Yuejin Xie, et al.
  arXiv 2026 (Core Contributor) | [Paper] [Code]

• Benchmarks for Trajectory Safety Evaluation and Diagnosis in OpenClaw and Codex: ATBench-Claw and ATBench-Codex
  Zhonghao Yang, Yu Li, Yanxu Zhu, Tianyi Zhou, Yuejin Xie, Haoyu Luo, Jing Shao, Xia Hu, Dongrui Liu
  arXiv 2026 | [Paper]

• Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability
  Qihan Ren, Peng Wang, Ruikun Cai, Shuai Shao, Dadi Guo, Yuejin Xie, Yafu Li, Quanshi Zhang, Xia Hu, Jing Shao, Dongrui Liu
  arXiv 2026 | [Paper]

• Frontier AI Risk Management Framework in Practice: A Risk Analysis Technical Report v1.5
  Dongrui Liu, …, Yuejin Xie, et al.
  arXiv 2026 | [Paper]

• AgentDoG: A Diagnostic Guardrail Framework for AI Agent Safety and Security
  Dongrui Liu, …, Yuejin Xie, et al. (Core Contributor)
  arXiv 2026 (Core Contributor) | [Paper] [Code]

• A Multi-Agent Conversational Bandit Approach to Online Evaluation and Selection of User-Aligned LLM Responses
  Xiangxiang Dai, Yuejin Xie, Maoli Liu, Xuchuang Wang, Zhuohua Li, Huanyu Wang, John C.S. Lui
  AAAI 2026 AI Alignment Track | [Paper] [arXiv]

• Towards Evaluating Proactive Risk Awareness of Multimodal Language Models
  Youliang Yuan, Wenxiang Jiao, Yuejin Xie, Chihao Shen, Menghan Tian, Wenxuan Wang, Jen-tse Huang, Pinjia He
  NeurIPS 2025 D&B Track | [Paper]

Education

• M.S., Tsinghua University (Shenzhen International Graduate School), Sep. 2026 - Jun. 2028 (expected), advised by Prof. Yujiu Yang
• B.E., Electronic Engineering, Huazhong University of Science and Technology, Sep. 2022 - Jun. 2026, GPA: 4.51/5.00, Rank: 1/30 (Advanced Class)

Internships

• Shanghai AI Lab, Dec. 2025 - Present
  Working with Dongrui Liu
  Research on LLM & Agent Safety

• The Chinese University of Hong Kong, Shenzhen, Jul. 2024 - Feb. 2025
  Working with Youliang Yuan, Dr. Wenxuan Wang and Prof. Pinjia He
  Research on LLM & Agent Safety

• The Chinese University of Hong Kong (Remote), Mar. 2024 - Jul. 2024
  Working with Xiangxiang Dai and Prof. John C.S. Lui
  Research on Multi-arms Bandit Algorithm

Awards

• National Scholarship (Top 1%), 2024
• National Scholarship (Top 1%), 2023
• The 6th DIGIX Global AI Challenge, National First Prize, 2024
• The 2024 Mathematical Contest in Modeling (MCM), Meritorious Winner, 2024