Ph.D. Student

Computer Science Department

University of Virginia

Email: [email protected] or [email protected]

Now I’m working on:

• LLM Reasoning

Twitter | Linkedin | CV

Zhihu | WeChat | Github

Google Scholar

Hi!👋 I’m Xinyu (/ʃɪn.ju:/) Zhu, ****a first-year computer science Ph.D. student at the University of Virginia, advised by Prof. ****Yu Meng.

I am broadly interested in NLP and ML, especially in solving complex reasoning problems with LLMs, improving generation quality and reducing hallucination of LLMs. My long-term research goal is to enable human expert level reasoning, decision-making and cognitive intelligence for neural models and systems.

I received my master degree at Tsinghua University, advised by Prof. ****Yujiu Yang. I have been a research intern at Microsoft Research Asia and International Digital Economy Academy.

✨News

• [2025.03] One survey on the honesty of LLMs is accepted by TMLR! 🎉
• [2025.01] One paper on evaluating LMM's reasoning via chart-to-code generation ****is accepted by ICLR 2025! 🎉
• [2024.10] Check out our comprehensive survey on the honesty of LLMs! 🎉 Also checkout our accompanied github repo! PRs welcome!
• [2024.09] One paper on improving MoE models' reasoning via self-contrast has been accepted by NeurIPS 2024, congrats to all the co-authors🎉! See you in Vancouver!
• [2024.09] Two papers on evaluating LLMs' hallucination under tool use setting and improving LLMs screenwriting via role-playing have been accepted by EMNLP 2024, congrats to all the co-authors🎉! See you in Miami!
• Old news

👨‍🎓Education

• University of Virginia 2024.08 -- Present Ph.D. in Computer Science
• Tsinghua University 2021.09 -- 2024.06 Master in Electronic and Information Engineering
• Xidian University 2017.09 -- 2021.06 Bachelor in Electronic Science and Technology

📝Selected Publications (* indicates equal contribution)

Unchosen Experts Can Contribute Too: Unleashing MoE Models' Power by Self-Contrast

NeurIPS 2024

Chufan Shi, Cheng Yang*, Xinyu Zhu*, Jiahao Wang*, Taiqiang Wu, Siheng Li, Deng Cai, Yujiu Yang, Yu Meng*

[paper][code]