Zhaoyi Li

Zhaoyi Li

PhD student in Computer Science

Hi:wave:, I am LI, ZHAOYI (Joey), a Ph.D. student at University of Science and Technology of China and City University of Hong Kong, co-adviced by Dr. Defu Lian (Situated Cognitive Intelligence Research Group @ USTC), Dr. Ying Wei (Compositional Lifelong Learning Lab @ Zhejiang University), and Dr. Linqi Song (Algorithmic Information theory and Artificial Intelligence Lab @ CityU). Previously, I got my Bachelor degree in Computer Science and Technology at University of Science and Technology of China in 2022.

My research goal is to build highly trustworthy and transparent language models that achieve human-like reasoning and systematic generalization. Specifically, I want to make human beings and neural network systems better understand each other (i.e., high interpretability and good generalization of blackbox deep models). I currently focus on investigating the compositionality of language models, which is about the mechanism that language models decompose complex questions and concepts into primitive ones and re-compose them step-by-step to grasp the complex semantic meaning. I am super curious about whether the end-to-end trained large-scale models are able to grok human-like reasoning skills or not and how to demonstrate it both empirically and theoretically. I would love to discuss on related topics should anyone is interested (always feel free to drop me an email: lizhaoyi777\(AT\)mail.ustc.edu.cn)!

I am currently focusing on the post-training of LLMs for enhancing their reasoning and generalization abilities and seeking opportunities of both industrial and academic research positions (wechat: joeylee2333).

πŸ”₯ News

  • 2026.01: Β πŸŽ– ICML 2026 Gold Reviewer Award.
  • 2026.01: Β πŸŽ‰πŸŽ‰ Two papers were accepted by ICLR 2026.
  • 2025.12: Β πŸŽ‰πŸŽ‰ I started an internship journey in Meituan LongCat Team.
  • 2025.08: Β πŸŽ‰πŸŽ‰ One paper was accepted by EMNLP 2025.
  • 2025.02: Β πŸŽ‰πŸŽ‰ One paper was accepted by ICLR 2025.
  • 2024.09: Β πŸŽ‰πŸŽ‰ One paper was accepted by Machine Learning Journal.
  • 2024.09: Β πŸŽ‰πŸŽ‰ One paper was accepted by EMNLP 2024.
  • 2024.05: Β πŸŽ‰πŸŽ‰ Two papers were accepted by ACL 2024.
  • 2023.05: Β πŸŽ‰πŸŽ‰ My first paper was accepted by ACL 2023.

πŸ“ Selected Papers

* indicates co-first authorship

Preprint 2026
On the Role of Reasoning Patterns in the Generalization Discrepancy of Long Chain-of-Thought Supervised Fine-Tuning
Zhaoyi Li*, Xiangyu Xi*, Zhengyu Chen, Wei Wang, Gangwei Jiang, Ranran Shen, Linqi Song, Ying Wei and Defu Lian
πŸ“„ Paper
ICLR 2026
Scaling Reasoning Hop Exposes Weaknesses: Demystifying and Improving Hop Generalization in Large Language Models
Zhaoyi Li, Jiatong Li, Gangwei Jiang, Linqi Song, Defu Lian and Ying Wei
πŸ“„ Paper πŸ’» Code
Preprint 2025
Learning to Substitute Components for Compositional Generalization
Zhaoyi Li, Gangwei Jiang, Chenwang Wu, Ying Wei, Defu Lian and Enhong Chen
πŸ“„ Paper
EMNLP 2024
Mitigating the Language Mismatch and Repetition Issues in LLM-based Machine Translation via Model Editing
Weichuan Wang*, Zhaoyi Li*, Defu Lian, Chen Ma, Linqi Song and Ying Wei
πŸ“„ Paper πŸ’» Code
ACL 2024 Findings
Understanding and Patching Compositional Reasoning in LLMs
Zhaoyi Li, Gangwei Jiang, Hong Xie, Linqi Song, Defu Lian and Ying Wei
πŸ“„ Paper πŸ’» Code πŸ“ Blog (δΈ­ζ–‡)
ACL 2024
Benchmarking and Improving Compositional Generalization of Multi-aspect Controllable Text Generation
Tianqi Zhong*, Zhaoyi Li*, Quan Wang, Linqi Song, Ying Wei, Defu Lian and Zhendong Mao
πŸ“„ Paper πŸ’» Code
ACL 2023 Oral Presentation
Learning to Substitute Spans towards Improving Compositional Generalization
Zhaoyi Li, Ying Wei and Defu Lian
πŸ“„ Paper πŸ’» Code

πŸ† Honors and Awards

  • 2026.05 ICML 2026 Gold Reviewer.
  • 2026.01 β€œDeChuang” USTC Scholarship for Overseas Visits for Excellent PhD Students.
  • 2024.10 China National Scholarship, USTC, Graduate Students.
  • 2022 ～ 2024 First prize, USTC Graduate Student Academic Scholarship.
  • 2022.06 Honored as The Outstanding Bachelor Graduate of USTC.
  • 2022.04 YuanQing (CEO of Lenovo Co.) Research Student Scholarship.
  • 2020 First prize, The 20th USTC RoboGame, Competition Group.
  • 2019 ～ 2022 First prize, USTC Excellent Undergraduate Scholarship.
  • 2019 USTC Ms.XIA Peisu Scholarship for Excellent Undergraduates of CS Major.
  • 2017.10 First prize, National (Senior) High School Mathematical Competition of China.
  • 2014.5 First prize, National (Junior) High School Mathematical Competition of China.

πŸ“– Educations

  • 2023.09 - now, PhD student, Department of Computer Science, City University of Hong Kong.
  • 2022.09 - now, PhD student, School of Computer Science and Technology, University of Science and Technology of China.
  • 2018.09 - 2022.07, Undergraduate, School of Computer Science and Technology, University of Science and Technology of China.

πŸ’» Internships

πŸŽ– Services

πŸˆβ€β¬› Special Links

  • Here is the link to the homepage of my girlfriend: Ranran Shen.