👋 Hi, I’m Yuqing! I’m a third-year Ph.D. student at the University of Southern California, advised by Prof. Robin Jia. Previously, I earned my M.S. at Fudan University with Prof. Xipeng Qiu and interned at GAIR Lab with Prof. Pengfei Liu.

I am broadly interested in building LLMs and LLM agents that improve themselves, that humans can trust, and that are efficient to train and run. My research currently spans three directions:

• 🔁 Self-improvement. How can LLMs and agents improve themselves without human supervision? I study how LLMs can surpass their weaker supervisors (weak-to-strong), and how agents can evolve their own memory (self-evolving memory extraction), tools (unsupervised tool evolution, paper coming soon), and harnesses.
• 🤝 Trustworthy collaboration. How can LLMs and agents become collaborators that humans can trust and oversee? I work on honesty, so that LLMs are upfront about what they know and admit their mistakes (honesty, retraction); memory, so that agents can better serve their users (memory); and transparency, so that humans can see what agents are actually doing.
• ⚡ Efficiency. How can LLMs and agents become faster and cheaper to train and run? I have worked on fine-tuning LLMs with limited resources (LOMO), and I’m increasingly interested in the time and cost of agents on long-horizon tasks.

News

• 🎤 [Oct 2026] I’ll be at COLM 2026 in San Francisco (Oct 5 to 9), giving an oral presentation on our retraction paper on the afternoon of Oct 6. Come say hi!
• ✈️ [Jul 2026] Attended ACL 2026 in San Diego, where I presented our table referencing paper as an oral, and our retraction paper won the Best Paper Award at KnowFM@ACL 2026.
• 📍 [May 2026] Joined Google, Sunnyvale (SVL) as a Student Researcher.

Education

• University of Southern California
  Ph.D. in Computer Science, 2024 - Present
  Advisor: Prof. Robin Jia
• Fudan University
  M.S. in Computer Science, 2021 - 2024
  Advisor: Prof. Xipeng Qiu
• University of Chinese Academy of Sciences
  B.E. in Computer Science, 2017 - 2021
  

Experience

• 🟦 Systems Research, Google
  Student Researcher, May 2026 - Aug. 2026
  Mentor: Jiani Zhang
  Worked on tool-centered evolution for data-lake analytics agents.
• 🟧 Amazon Bedrock Core Science
  Applied Scientist Intern, May 2025 - Aug. 2025
  Mentor: Qi Zhu
  Enhanced LLM robustness for table tasks by identifying failure modes and optimizing data referencing accuracy.

Selected Publications

* denotes co-first authors

Self-Evolving LLM Memory Extraction Across Heterogeneous Tasks
Yuqing Yang, Tengxiao Liu, Wang Bill Zhu, Taiwei Shi, Linxin Song, Robin Jia
Preprint 2026 [paper] [code]

Can Coding Agents Optimize Algorithms Autonomously?
Tengxiao Liu, Yuqing Yang, Xi Ye, Danqi Chen
Blog 2026 [blog] [code]

When LLMs Read Tables Carelessly: Measuring and Reducing Data Referencing Errors
Yuqing Yang, Qi Zhu, Zhen Han, Boran Han, Zhengyuan Shen, Shuai Wang, Vassilis N. Ioannidis, Huzefa Rangwala
ACL 2026 Oral [paper] [code]

When Do LLMs Admit Their Mistakes? Understanding the Role of Model Belief in Retraction
Yuqing Yang, Robin Jia
CoLM 2026 Oral 🏆 KnowFM@ACL 2026 Best Paper [paper] [code]

Weak-to-Strong Reasoning
Yuqing Yang, Yan Ma, Pengfei Liu
EMNLP Findings 2024 [paper] [code]

Alignment for Honesty
Yuqing Yang, Ethan Chern, Xipeng Qiu, Graham Neubig, Pengfei Liu
NeurIPS 2024 [paper] [code]

Full Parameter Fine-tuning for Large Language Models with Limited Resources
Kai Lv, Yuqing Yang, Tengxiao Liu, Qinghui Gao, Qipeng Guo, Xipeng Qiu
ACL 2024 Oral [paper] [code]

Thank you!