Jianli Zhao

pic.jpg

zhao78740 AT gmail.com

Hi, there.

I am Jianli Zhao, who earned a B.S. in Computer Science from Shandong University, where I was mentored by Prof. Bin Jiang. Now I am spending a serene gap year, during which I have the privilege of being mentored by Prof. Fazl Barez from Oxford. Earlier, in the summer of 2024, I also spent a golden summer at UCSD, working under the guidance of Prof. Julian Mcauley.

My prior research primarily concentrated on Information Extraction using a diffusion model method [PDF]. Currently I am keen on Trustworthy LLMs and AI safety. My interests span from exploring the vulnerabilities of LLMs, such as red-teaming, to investigating interpretability via circuit-level analysis, with the goal of ensuring a more secure deployment. [PDF]


Research Interest

  • Natural Language Processing
  • AI Safety
  • Trustworthy LLMs
  • Interpretability and Alignment


Selected Publications

  • Chain-of-Thought Hijacking
    Jianli Zhao, Tingchen Fu, Rylan Schaeffer, Mrinank Sharma, Fazl Barez
    Under Review 2025   [paper]   [code]   [website]

  • IPED: An Implicit Perspective for Relational Triple Extraction based on Diffusion Model
    Jianli Zhao, Changhao Xu, Bin Jiang
    NAACL 2024   [paper]   [code]


Education

  • B.S. in Shandong University, 2021-2025