Wenjie Jacky Mo
I am a first-year Ph.D. student in Computer Science at the University of California, Davis, where I am fortunate to be advised by Prof. Muhao Chen and Prof. Zhe Zhao.
My research primarily focuses on AI safety, particularly in building secure and reliable LLMs and LLM4Code. I explore various aspects of safety and robustness of these models, including defending against and detecting backdoor attacks. In addition, my work extends to understanding and mitigating jailbreak attacks, where models are manipulated to generate harmful or unintended outputs.
news
| Jun 25, 2025 | Excited to share that our new paper, RedCoder: Automated Multi‑Turn Red Teaming for Code LLMs is now available on arXiv! It introduces RedCoder, an autonomous agent that systematically generates multi-turn adversarial interactions to expose vulnerabilities in code-generating LLMs. |
|---|---|
| May 15, 2025 | Excited to share that this summer I’ll be joining the Center for AI Safety as a Research Intern! My work will focus on evaluating LLM behavior in legal tort contexts and exploring multimodal utility stimuli for model analysis. |
| Sep 23, 2024 | Thrilled to share that our team, CapitalAI, has been selected as a Red-Team competitor in the Amazon Trust AI Challenge! Honored to serve as the Co-lead of our team as we tackle this exciting opportunity. |