My name is Jiashu Qu, and I am a first-year master student at Peking University, advised by Prof. Wentao Zhang.

During my undergraduate studies, I was fortunate to work with Prof. Yue Zhao at USC and Dr. Hongcheng Gao at Tsinghua University.

My research interests broadly center on alignment in systems built upon large models, including LLMs, MLLMs, VLA, and Agent-based systems. I am interested in both alignment failures across modalities (e.g. hallucination) and misalignment with human objectives in real-world interactive systems.

📝 Research

• SafeWork-R1: Coevolving Safety and Intelligence under the AI-45 Law
  (Core Contributor) Shanghai AI Lab et al.
  Tech Report
  📄Paper 🤗HuggingFace

• Exploring Hallucination of Large Multimodal Models in Video Understanding: Benchmark, Analysis and Mitigation
  H. Gao*, J. Qu*, J. Tang*, B. Bi, Y. Liu, H. Chen, L. Liang, L. Su, Q. Huang
  Transaction of ACL
  📄Paper 🤗HuggingFace

• Treble Counterfactual VLMs: A Causal Approach to Hallucination
  S. Li, J. Qu, Y. Zhou, Y. Qin, T. Yang, Y. Zhao
  EMNLP 2025
  📄Paper

• The Devil Behind the Mask: An Emergent Safety Vulnerability of Diffusion LLMs
  Z. Wen, J. Qu, D. Liu, Z. Liu, R. Wu, Y. Yang, X. Jin, H. Xu, X. Liu, W. Li, C. Lu
  ICLR 2026
  📄Paper

• SafeDriveVLA: Navigation-Conditioned World Model Dreaming for Conflict-Aware End-to-End Autonomous Driving
  S. Xie*, Z. Zhang*, J. Wang, J. Qu, X. Liang, L. Kong, J. Lu, H. I. Christensen, Q. A. Chen
  CoRL 2026
  📄Paper

• Read, Route, Patch: Robust Test-Time Evolution for LLM Agents
  K. Chen, L. Yan, Y. Wei, J. Luo, Z. Dou, J. Qu, H. Liu, J. Wang
  NeurIPS 2026 Workshop on Towards Test-Time Continual Learning Agents
  📄Paper

💻 Experience

• Algorithm Intern, ByteDance
  May 2026 – July 2026
  

• Research Intern, The Center of Safe & Trustworthy AI @ Shanghai AI Lab
  May 2025 – Aug 2025
  Mentor: Dongrui Liu

• Research Intern, FORTIS Lab @ University of Southern California
  Dec 2024 – Sep 2025
  Mentor: Yue Zhao

📖 Educations

• 2021.09 - 2026.05, Chongqing University & University of Cincinnati
  • Rank: 1 / 97

🧑‍💻 Academic Service

Reviewer for

• 2027: ICLR
• 2026: EMNLP, ESORICS

🏆 Honors and Awards

• National Scholarship (Top 1, twice)
• Meritorious Winner, Mathematical Contest in Modeling (MCM/ICM)

👀 Visitors: · 📊 Page Views: