About

Haibo Jin is a PhD student at the University of Illinois Urbana-Champaign, advised by Prof. Haohan Wang in the DREAM Lab. His research focuses on building reliable, efficient, and scalable agentic AI systems through reusable primitives and harness engineering. His broader interests include AI safety and generative engine optimization (GEO), with an emphasis on improving the trustworthiness and effectiveness of AI systems in real-world applications.

Research interests

• AI Safety
• Reusable Primitives
• Agentic AI Systems
• Generative Engine Optimization

Highlight

Publications

1. Agent Primitives: Reusable Latent Building Blocks for Multi-Agent Systems. Haibo Jin, Peng Kuang, Ye Yu, Xiaopeng Yuan, Haohan Wang. ICML 2026. arXiv Project page
2. Evaluating the Inductive Abilities of Large Language Models: Why Chain-of-Thought Reasoning Sometimes Hurts More Than Helps. Haibo Jin, Peiyan Zhang, Man Luo, Haohan Wang. NeurIPS 2025. OpenReview
3. Exploring the Vulnerability of the Content Moderation Guardrail in Large Language Models via Intent Manipulation. Jun Zhuang, Haibo Jin, Ye Zhang, Zhengjian Kang, Wenbin Zhang, Gaby G. Dagher, Haohan Wang. Findings of EMNLP 2025. ACL Anthology
4. REVOLVE: Optimizing AI Systems by Tracking Response Evolution in Textual Optimization. Peiyan Zhang, Haibo Jin, Leyang Hu, Xinnuo Li, Liying Kang, Man Luo, Yangqiu Song, Haohan Wang. ICML 2025. arXiv
5. Jailbreaking Large Language Models Against Moderation Guardrails via Cipher Characters. Haibo Jin, Andy Zhou, Joe D. Menke, Haohan Wang. NeurIPS 2024. arXiv
6. CatchBackdoor: Backdoor Detection via Critical Trojan Neural Path Fuzzing. Haibo Jin, Ruoxi Chen, Jinyin Chen, Haibin Zheng, Yang Zhang, Haohan Wang. ECCV 2024. DOI
7. EditShield: Protecting Unauthorized Image Editing by Instruction-guided Diffusion Models. Ruoxi Chen, Haibo Jin, Yixin Liu, Jinyin Chen, Haohan Wang, Lichao Sun. ECCV 2024. DOI
8. GUARD: Role-playing to Generate Natural-language Jailbreakings to Test Guideline Adherence of Large Language Models. Haibo Jin, Ruoxi Chen, Andy Zhou, Yang Zhang, Haohan Wang. Workshop on Secure and Trustworthy Large Language Models (SeT LLM @ ICLR 2024), 2024. arXiv
9. Fight Perturbations with Perturbations: Defending Adversarial Attacks via Neuron Influence. Ruoxi Chen, Haibo Jin, Haibin Zheng, Jinyin Chen. IEEE Transactions on Dependable and Secure Computing (TDSC), 2024. DOI
10. AdvCheck: Characterizing adversarial examples via local gradient checking. Ruoxi Chen, Haibo Jin, Jinyin Chen, Haibin Zheng, Shilian Zheng, Xiaoniu Yang, Xing Yang. Computers & Security, 136: 103540, 2024. DOI
11. Deep Learning Testing for Reliability: A Survey (in Chinese). Ruoxi Chen, Haibo Jin, Jinyin Chen, Haibin Zheng, Xiaohao Li. Journal of Cyber Security, 9(1), 2024.
12. CertPri: Certifiable Prioritization for Deep Neural Networks via Movement Cost in Feature Space. Haibin Zheng, Jinyin Chen, Haibo Jin. 38th IEEE/ACM International Conference on Automated Software Engineering (ASE), 2023. DOI
13. Excitement surfeited turns to errors: Deep learning testing framework based on excitable neurons. Haibo Jin, Ruoxi Chen, Haibin Zheng, Jinyin Chen, Yao Cheng, Yue Yu, Tieming Chen, Xianglong Liu. Information Sciences, 637: 118936, 2023. DOI
14. ROBY: Evaluating the adversarial robustness of a deep model by its decision boundaries. Haibo Jin, Jinyin Chen, Haibin Zheng, Zhen Wang, Jun Xiao, Shanqing Yu, Zhaoyan Ming. Information Sciences, 587: 97–122, 2022. DOI
15. Prioritization Testing for Infrared Object Detection based on Significant Feature Mutation. Haibin Zheng, Yunjie Yan, Jinyin Chen, Haibo Jin, Ruoxi Chen, Qi Xuan. Earth & Space: From Infrared to Terahertz International Conference (ESIT 2022), 2022. Oral.

For a complete and up-to-date list, see Google Scholar.

Education

• Ph.D. in Information Sciences, University of Illinois Urbana-Champaign · 2025–present
• Ph.D. program in Control Science and Engineering, Zhejiang University of Technology · 2022–2024
• M.Eng. in Control Science and Engineering, Zhejiang University of Technology · 2020–2022
• B.Eng. in Communication Engineering, Zhejiang University of Technology · 2016–2020

Research experience

• Research Scholar, University of Illinois Urbana-Champaign, with Prof. Haohan Wang · from March 2023
  Security testing and robustness of large language models: jailbreak and testing methods for LLMs, prioritization of unlabeled test cases, and testing vision-language models under adversarial and backdoor attacks.
• Research Assistant, Zhejiang University of Technology, with Prof. Jinyin Chen · September 2020–March 2023
  Evasion attacks and defenses for deep learning systems: defense and detection methods against adversarial attacks, robustness evaluation of deep learning models, and threats in model supply chains.

Selected awards

• First Prize, China Collegiate Computing Contest, Artificial Intelligence Innovation Contest · 2020
• Postgraduate Premium Scholarship, Zhejiang University of Technology · 2021
• Excellent Postgraduate, Zhejiang University of Technology · 2021
• “Top 10 Academic Stars” Scholarship, Zhejiang University of Technology · 2022
• Nominee Award, 8th China Postgraduate Smart City Technology and Creative Design Competition (“West Lake Scientific Innovation Cup”) · 2022