publications

the first authors with * contributed equally

2026

  1. ACL 2026
    Reinforcement learning for self-improving agent with skill library
    Jiongxiao Wang, Qiaojing Yan, Yawei Wang, Yijun Tian, Soumya Smruti Mishra, Zhichao Xu, Megha Gandhi, Panpan Xu, and Lin Lee Cheong
    In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics, 2026

2025

  1. ICLR 2025
    Robust Representation Consistency Model via Contrastive Denoising
    Julius Berner, Jiongxiao Wang, Zhongzhu Chen, Chaowei Xiao, Zhongjie Ba, Kui Ren, Jun Zhu, and Anima Anandkumar
    In International Conference on Learning Representations, 2025
  2. ICLR 2025
    Benchmarking vision language model unlearning via fictitious facial identity dataset
    Yingzi Ma, Jiongxiao Wang, Fei Wang, Siyuan Ma, Jiazhao Li, Jinsheng Pan, Xiujun Li, Furong Huang, Lichao Sun, Bo Li, and 3 more authors
    International Conference on Learning Representations, 2025
  3. IEEE S&P
    Preference Poisoning Attacks on Reward Model Learning
    Junlin Wu, Jiongxiao Wang, Chaowei Xiao, Chenguang Wang, Ning Zhang, and Yevgeniy Vorobeychik
    IEEE S&P, 2025
  4. NAACL 2025
    Test-time backdoor mitigation for black-box large language models with defensive demonstrations
    Wenjie Mo, Jiashu Xu, Qin Liu, Jiongxiao Wang, Jun Yan, Hadi Askari, Chaowei Xiao, and Muhao Chen
    In Findings of the Association for Computational Linguistics, 2025

2024

  1. Preprint
    Fath: Authentication-based test-time defense against indirect prompt injection attacks
    Jiongxiao Wang, Fangzhou Wu, Wendi Li, Jinsheng Pan, Edward Suh, Z Morley Mao, Muhao Chen, and Chaowei Xiao
    arXiv preprint arXiv:2410.21492, 2024
  2. Preprint
    Safeguarding Vision-Language Models Against Patched Visual Prompt Injectors
    Jiachen Sun, Changsheng Wang, Jiongxiao Wang, Yiwei Zhang, and Chaowei Xiao
    arXiv preprint arXiv:2405.10529, 2024
  3. NeurIPS 2024
    Mitigating Fine-tuning based Jailbreak Attack with Backdoor Enhanced Safety Alignment
    Jiongxiao Wang, Jiazhao Li, Yiquan Li, Xiangyu Qi, Muhao Chen, Junjie Hu, Yixuan Li, Bo Li, and Chaowei Xiao
    Thirty-Eighth Annual Conference on Neural Information Processing Systems, 2024
  4. NeurIPS 2024
    Consistency Purification: Effective and Efficient Diffusion Purification towards Certified Robustness
    Yiquan Li*, Zhongzhu Chen*, Kun Jin*, Jiongxiao Wang*, Bo Li, and Chaowei Xiao
    Thirty-Eighth Annual Conference on Neural Information Processing Systems, 2024
  5. ACL 2024
    RLHFPoison: Reward Poisoning Attack for Reinforcement Learning with Human Feedback in Large Language Models
    Jiongxiao Wang, Junlin Wu, Muhao Chen, Yevgeniy Vorobeychik, and Chaowei Xiao
    In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics, 2024
  6. ICLR 2024
    Conversational Drug Editing Using Retrieval and Domain Feedback
    Shengchao Liu, Jiongxiao Wang, Yijin Yang, Chengpeng Wang, Ling Liu, Hongyu Guo, and Chaowei Xiao
    In The Twelfth International Conference on Learning Representations, 2024

2023

  1. Preprint
    Adversarial Demonstration Attacks on Large Language Models
    Jiongxiao Wang*, Zichen Liu*, Keun Hee Park, Muhao Chen, and Chaowei Xiao
    arXiv preprint arXiv:2305.14950, 2023
  2. NeurIPS 2023
    On the exploitability of instruction tuning
    Manli Shu, Jiongxiao Wang, Chen Zhu, Jonas Geiping, Chaowei Xiao, and Tom Goldstein
    Advances in Neural Information Processing Systems, 2023
  3. ICML 2023
    A Critical Revisit of Adversarial Robustness in 3D Point Cloud Recognition with Diffusion-Driven Purification
    Jiachen Sun, Jiongxiao Wang, Weili Nie, Zhiding Yu, Zhuoqing Mao, and Chaowei Xiao
    In Proceedings of the 40th International Conference on Machine Learning, 2023

2022

  1. ICLR 2022
    Densepure: Understanding diffusion models for adversarial robustness
    Chaowei Xiao*, Zhongzhu Chen*, Kun Jin*, Jiongxiao Wang*, Weili Nie, Mingyan Liu, Anima Anandkumar, Bo Li, and Dawn Song
    In The Eleventh International Conference on Learning Representations, 2022
  2. ICLR 2022
    Defending against Adversarial Audio via Diffusion Model
    Shutong Wu, Jiongxiao Wang, Wei Ping, Weili Nie, and Chaowei Xiao
    In The Eleventh International Conference on Learning Representations, 2022
  3. ICML 2022
    Fast and reliable evaluation of adversarial robustness with minimum-margin attack
    Ruize Gao, Jiongxiao Wang, Kaiwen Zhou, Feng Liu, Binghui Xie, Gang Niu, Bo Han, and James Cheng
    In International Conference on Machine Learning, 2022