About Me

Luke Yukun Jiang (江钰坤 in Chinese)

I am Luke, a Ph.D. student at CISPA Helmholtz Center for Information Security (Saarbrücken, Germany), advised by Prof. Michael Backes and Dr. Yang Zhang. I obtained my B.E. degree at Sichuan University (Chengdu, China) advised by Prof. Beibei Li in 2022. My research interests broadly span AI Security.

Apr. 2026: our work ‘‘Sparse Models, Sparse Safety: Unsafe Routes in Mixture-of-Experts LLMs’’ has been accepted by ICML 2026. See you in Seoul.
Apr. 2026: two of our works (‘‘Open Schrödinger’s Closed Box: Identifying Retrieval Augmented Generation in API-Accessible Large Language Model Services’’ and ‘‘DE-CLIP: Few-Shot Anomaly Detection via Difference-Guided Embedding Editing’’) have been accepted by ACL 2026. See you in San Diego.
Mar. 2026: our work ‘‘Real Money, Fake Models: Deceptive Model Claims in Shadow APIs’’ has been covered by [JIQIZHIXIN | Synced (Chinese)].
Feb. 2026: our work ‘‘“Humans welcome to observe”: A First Look at the Agent Social Network Moltbook 🦞’’ has been covered by [TechXplore] and [AI Era (Chinese)].
Sep. 2025: our work ‘‘Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency’’ has been accepted by NeurIPS 2025. See you in San Diego.

[arXiv] Yukun Jiang, Yage Zhang, Michael Backes, Xinyue Shen, Yang Zhang: ‘‘HarmfulSkillBench: How Do Harmful Skills Weaponize Your Agents?’’ [arXiv] [Code] [Dataset]
[arXiv] Yukun Jiang*, Yage Zhang*, Xinyue Shen*, Michael Backes, Yang Zhang: ‘‘“Humans welcome to observe”: A First Look at the Agent Social Network Moltbook.’’ [Website] [PDF] [Dataset] [arXiv] Media Coverage: [TechXplore] [AI Era (Chinese)]
[ICML 2026] Yukun Jiang, Hai Huang, Mingjie Li, Yage Zhang, Michael Backes, Yang Zhang: ‘‘Sparse Models, Sparse Safety: Unsafe Routes in Mixture-of-Experts LLMs.’’ [arXiv] [Code]
[ACL 2026] Yukun Jiang, Xinyue Shen, Michael Backes, Zheng Li, Yang Zhang: ‘‘Open Schrödinger’s Closed Box: Identifying Retrieval Augmented Generation in API-Accessible Large Language Model Services.’’
[NeurIPS 2025] Yukun Jiang, Mingjie Li, Michael Backes, Yang Zhang: ‘‘Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency.’’ [PDF] [Link] [Code]
[EMNLP 2024] Yukun Jiang, Zheng Li, Xinyue Shen, Yugeng Liu, Michael Backes, Yang Zhang: ‘‘ModSCAN: Measuring Stereotypical Bias in Large Vision-Language Models from Vision and Language Modalities.’’ [PDF] [Link] [Code]
[ICWSM 2024] Yukun Jiang, Xinyue Shen, Rui Wen, Zeyang Sha, Junjie Chu, Yugeng Liu, Michael Backes, Yang Zhang: ‘‘Games and Beyond: Analyzing the Bullet Chats of Esports Livestreaming.’’ [PDF] [Link]

— “What I’ve done cannot be undone.”