Xiaoyan Bai

PhD student, Department of Computer Science, University of Chicago

Joined

January 2024

Names

Xiaoyan Bai (Preferred)

Emails

****@umich.edu (Confirmed)

****@uchicago.edu (Confirmed)

Personal Links

Homepage

Google Scholar

DBLP

Career & Education History

PhD student

Department of Computer Science, University of Chicago (uchicago.edu)

2024 – Present

Undergrad student

College of Engineering, University of Michigan - Ann Arbor (umich.edu)

2022 – 2024

Advisors, Relations & Conflicts

PhD Advisor

Chenhao Tan

2024 – Present

Expertise

Computational Social Science

2024 – Present

Mechanistic Interpretability

2023 – Present

Publications

Agent-Native Research Artifacts
Jiachen Liu, Jiaxin Pei, Jintao Huang, Chenglei Si, Ao Qu, Xiangru Tang, Runyu Lu, Xiaoyan Bai, Haizhong Zheng, Zhiyang Chen, Zexue He, Haojie Ye, Yujuan Fu, Zijian Jin, Zhenyu Zhang, Shangquan Sun, Dianzhuo Wang, Qian-Ze Zhu, Jiachen Sun, Mingyuan Wu et al. (14 additional authors not shown)
- AgentSkills 2026 Oral
- Readers: Everyone
Know Thyself? On the Incapability and Implications of AI Self-Recognition
Xiaoyan Bai, Aryan Shrivastava, Ari Holtzman, Chenhao Tan
- MSLD 2026 Poster
- Readers: Everyone
The Story is Not the Science: Execution-Grounded Evaluation of Mechanistic Interpretability Research
Xiaoyan Bai, Alexander Baumgartner, Haojia Sun, Ari Holtzman, Chenhao Tan
- MSLD 2026 Poster
- Readers: Everyone
Know Thyself? On the Incapability and Implications of AI Self-Recognition
Xiaoyan Bai, Aryan Shrivastava, Ari Holtzman, Chenhao Tan
- Submitted to ICLR 2026
- Readers: Everyone
Why Can't Transformers Learn Multiplication? Reverse-Engineering Implicit Chain-of-Thought Reveals Challenges of Learning Long-Range Dependencies
Xiaoyan Bai, Itamar Pres, Yuntian Deng, Chenhao Tan, Stuart Shieber, Fernanda Viégas, Martin Wattenberg, Andrew Lee
- Submitted to ICLR 2026
- Readers: Everyone
``You are a brilliant mathematician'' Does Not Make LLMs Act Like One
Xiaoyan Bai, Ari Holtzman, Chenhao Tan
- Agents4Science
- Readers: Everyone
Concept Incongruence: An Exploration of Time and Death in Role Playing
Xiaoyan Bai, Ike Peng, Aditya Singh, Chenhao Tan
- NeurIPS 2025 poster
- Readers: Everyone
Learn To be Efficient: Build Structured Sparsity in Large Language Models
Haizhong Zheng, Xiaoyan Bai, Xueshen Liu, Zhuoqing Mao, Beidi Chen, Fan Lai, Atul Prakash
- NeurIPS 2024 spotlight
- Readers: Everyone
A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity
Andrew Lee, Xiaoyan Bai, Itamar Pres, Martin Wattenberg, Jonathan K. Kummerfeld, Rada Mihalcea
- ICML 2024 Oral
- Readers: Everyone
Learn To be Efficient: Build Structured Sparsity in Large Language Models
Haizhong Zheng, Xiaoyan Bai, Beidi Chen, Fan Lai, Atul Prakash
- CoRR 2024
- Readers: Everyone

View all 13 publications

Co-Authors

View all 52 co-authors