Zhiyuan Zhong

me.jpg

I am currently a CS PhD student at University of Utah, advised by Prof. Guanhong Tao. I obtained my B.E. in Computer Science from SUSTech, advised by Prof. Yepang Liu. I also spend a wonderful time as a research assistant at HKUST(GZ), advised by Prof. Xinlei He.

My research interests lie in LLM4Code and LLM/VLM safety & security. I’ve helped identify over 30 bugs in fastjson2, a popular open source JSON library by Alibaba.

News

May, 2025 MGT-Academic (benchmarking AIGT detection) got accepted in KDD D&B track 2025.
Oct, 2024 Check out MGTBench2.0, a comprehensive framework for machine-generated text detection, and AITextDetect, a dataset spanning topics like STEM with content from 5 popular LLMs.
Jul, 2024 Outstanding Performance Award of Information Hub Summer Camp at HKUST(GZ).

Publications

  1. Arxiv
    Backdoor Attack on Vision Language Models with Stealthy Semantic Manipulation
    Zhiyuan Zhong, Zhen Sun, Yepang Liu, Xinlei He, and Guanhong Tao
    Preprint, 2025
    PDF
  2. KDD 2025 D&B
    On the Generalization and Adaptation Ability of Machine-Generated Text Detectors in Academic Writing
    Yule Liu*, Zhiyuan Zhong*, Yifan Liao, Zhen Sun, Jingyi Zheng, Jiaheng Wei, Qingyuan Gong, Fenghua Tong, Yang Chen, Yang Zhang, and Xinlei He
    ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2025
    PDF
  3. Arxiv
    Advancing Bug Detection in Fastjson2 with Large Language Models Driven Unit Test Generation
    Zhiyuan Zhong, Sinan Wang, Hailong Wang, Shaojin Wen, Hao Guan , Yida Tao, and Yepang Liu
    Preprint, 2024
    PDF