Yingqian Cui's Homepage
I am a Ph.D. student in the Department of Computer Science and Engineering at Michigan State University. I am co-advised by Prof. Jiliang Tang and Prof. Yue Xing. I received my B.Sc. (Honors) in Mathematics and Statistics from Hong Kong Baptist University.
My research focuses on the reasoning capabilities of large language models, with interests spanning efficient and effective reasoning, theoretical understanding of in-context learning and chain-of-thought, and trustworthy AI.
Research Interests
- Reasoning of Large Language Models
- Trustworthy AI: Safety, Robustness, and Fairness
- Theory in Deep Learning
Selected Updates
- Completed my Applied Scientist internship at Microsoft.
- Our work "** Important** You should give me full credits!": Exploring Prompt Injection Attacks on LLM-Based Automatic Grading Systems will appear in Findings of EMNLP 2026.
- New submission: Adaptive Multi-Step Lookahead Decoding for Diffusion Language Models.
- Our work Adaptive Test-Time Reasoning via Reward-Guided Dual-Phase Search appeared in Findings of ACL 2026.
- Our work Retrieval Heads are Dynamic appeared at ACL 2026.
Selected Publications
Findings of ACL, 2026
ACL, 2026
Stepwise Perplexity-Guided Refinement for Efficient Chain-of-Thought Reasoning in Large Language Models
Findings of ACL, 2025
A Theoretical Understanding of Chain-of-Thought: Coherent Reasoning and Error-Aware Demonstration
AISTATS, 2025