1 short paper accepted by EMNLP 2025; 2 papers accepted by COLM 2025; 1 paper accepted by ICML 2025; 1 paper accepted by ICLR 2025; 1 paper accepted by TMLR 2024.
Research Experience
No detailed information provided.
Education
PhD: Department of Computer Science at Purdue University, Advisor: Ruqi Zhang; BE: Computer Science and Technology at Tianjin University, Advisor: Changqing Zhang.
Background
Research Interests: Building statistical frameworks to enhance the stability and efficiency of LLM post-training. Recently focusing on the sampling process in RLVR algorithms. Also broadly interested in preference alignment, (multimodal) LLM safety, and Bayesian deep learning.