I am an incoming Ph.D. student in Computer Science at the College of Computer Science and Technology, Zhejiang University, starting in Fall 2027, advised by Prof. Bohan Zhuang. I am currently an undergraduate student at the School of Artificial Intelligence, Nanjing University.
My research interests focus on embodied agents and efficient embodied model architectures. I have previously explored LLM post-training, LLM agents, and VLA/WM extensively.
Publications
ToMap refines proof decompositions at test time to improve multi-agent Lean proof autoformalization.
Token-wise clipping during SFT preserves prior knowledge and provides a more generalizable initialization for reinforcement learning.
An L0–L7 evaluation ladder connects world-model evidence to embodied decision-making, from visual plausibility to policy optimization.
Education
School of Artificial Intelligence
Research Intern at LAMDA RL LAB, advised by Prof. Yang Yu
Awards
- National Scholarship of China
- National Special Prize in the Main Competition of the Challenge Cup (one of the seven reporting projects at the closing ceremony)
- Nanjing University Science and Technology Innovation Scholarship
- International First Prize in the Mathematical Contest in Modeling (MCM/ICM)
- GeoX Interdisciplinary Project for Young Students, Frontiers Science Center for Critical Earth Material Cycling