About Me

I’m an incoming M.Phil. student in Computer Science at the University of Cambridge, and a member of Hughes Hall. Prior to this, I spent several years as a research intern at CASIA, where I was fortunate to be advised by Prof. Shu Wu and work with Dr. Mengqi Zhang.

My work focuses on understanding and generalizably controlling large language models (LLMs), particularly how a (relatively) small number of samples can enable truly generalizable knowledge updating, capability evolution, value alignment, and behavioral control. I am likewise interested in closely related topics such as interpretability, evaluation, and safety issues, with the ultimate goal of making LLMs more robust and trustworthy for real-world deployment, thereby empowering diverse domains and benefiting society at large.

Recent Update

• [Apr. 2026] Spectral Characterization and Mitigation of Sequential Knowledge Editing Collapse gets accepted to ACL 2026!
• [Jan. 2026] Disentangling Knowledge Representations for Large Language Model Editing gets accepted to ICLR 2026!
• [Oct. 2025] Thrilled to be awarded the CCF Elite Collegiate Award! Heading to CNCC2025@Harbin for the ceremony!
• [Sept. 2025] LLM Unlearning Should Be Form-Independent gets accepted to S&P 2026!! See you in San Francisco!
• [Aug. 2025] Two papers get accepted to EMNLP 2025!
• [Feb. 2025] Uncovering Overfitting in Large Language Model Editing gets accepted to ICLR 2025 as Spotlight! See you in Singapore!

Research Highlights