Yafu Li (李雅夫)

Selected Publications

My research focuses on reasoning, trustworthy AI, and multilinguality. * denotes equal contribution; † denotes project lead or corresponding author.

Reasoning and RL

Test-Time Preference Optimization: On-the-Fly Alignment via Iterative Textual Feedback

Yafu Li, Xuyang Hu, Xiaoye Qu, Linjie Li, Yu Cheng

ICML, 2025

Code / Paper

Aligning LLMs at test time via textual reward.

Learning to Reason under Off-Policy Guidance

Jianhao Yan^*, Yafu Li^*†, Zican Hu, Zhi Wang, Ganqu Cui, Xiaoye Qu, Yu Cheng, Yue Zhang

NeurIPS, 2025

Code / Paper

Boosting reasoning performance using off-policy guidance.

ExGRPO: Learning to Reason from Experience

Runzhe Zhan, Yafu Li^†, Zhi Wang, Xiaoye Qu, Dongrui Liu, Jing Shao, Derek F. Wong, Yu Cheng

ICLR, 2026

Code / Paper

Boosting reasoning performance with the model's own experience.

Diversity-Incentivized Exploration for Versatile Reasoning

Zican Hu, Shilin Zhang, Yafu Li^†, Jianhao Yan, Xuyang Hu, Leyang Cui, Xiaoye Qu, Chunlin Chen, Yu Cheng, Zhi Wang

ICLR, 2026

Code / Paper

Encouraging global exploration.

Conditional Advantage Estimation for Reinforcement Learning in Large Reasoning Models

Guanxu Chen, Yafu Li^†, Yuxian Jiang, Chen Qian, Qihan Ren, Jingyi Yang, Yu Cheng, Dongrui Liu, Jing Shao

ICLR, 2026

Code / Paper

Digging implicit signals via group comparison.

A Survey of Efficient Reasoning for Large Reasoning Models: Language, Multimodality, and Beyond

Xiaoye Qu^*†, Yafu Li^*†, Zhaochen Su, Weigao Sun, Jianhao Yan, Dongrui Liu, Ganqu Cui, Daizong Liu, Shuxian Liang, Junxian He, Peng Li, Wei Wei, Jing Shao, Chaochao Lu, Yue Zhang, Xian-Sheng Hua, Bowen Zhou, Yu Cheng

preprint

Code / Paper

A survey of efficient reasoning for Large Reasoning Models.

Characterizing, Evaluating, and Optimizing Complex Reasoning

Haoran Zhang, Yafu Li^†, Zhi Wang, Zhilin Wang, Shunkai Zhang, Xiaoye Qu, Yu Cheng

preprint

Code / Paper

A framework for robust and scalable evaluation for complex reasoning.

Scaling Reasoning, Losing Control: Evaluating Instruction Following in Large Reasoning Models

Tingchen Fu, Yafu Li^†, Jiawei Gu, Xiaoye Qu, Yu Cheng

ACL, 2026

Code / Paper

A tension between scaling up reasoning capacity and maintaining controllability.

From Drafts to Answers: Unlocking LLM Potential via Aggregation Fine-Tuning

Yafu Li^*, Zhilin Wang^*, Tingchen Fu, Ganqu Cui, Sen Yang, Yu Cheng

ACL Findings, 2026

Code / Paper

Training models to aggregate multiple responses.

Multi-LLM Collaborative Search for Complex Problem Solving

Sen Yang, Yafu Li^†, Wai Lam, Yu Cheng

ACL Findings, 2026

Paper

Aggregate multiple agents for complex reasoning.

Trustworthy AI

Reasoning over Boundaries: Enhancing Specification Alignment via Test-time Deliberation

Haoran Zhang, Yafu Li^†, Xuyang Hu, Dongrui Liu, Zhilin Wang, Bo Li, Yu Cheng

preprint

Code / Paper

Reasoning over safety and behavioural boundaries before answering.

Unveiling Attractor Cycles in Large Language Models: A Dynamical Systems View of Successive Paraphrasing

Zhilin Wang^*, Yafu Li^*†, Jianhao Yan, Yu Cheng, Yue Zhang

ACL, 2025

Paper

Unveiling attractor cycles in LLMs through dynamical systems analysis.

MAGE: Machine-generated Text Detection in the Wild

Yafu Li, Qintong Li, Leyang Cui, Wei Bi, Longyue Wang, Linyi Yang, Shuming Shi, Yue Zhang

ACL, 2024

Code / Paper

Assessing machine-generated text detectors in real-world scenarios.

Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Models

Yue Zhang, Yafu Li, Leyang Cui, Deng Cai, Lemao Liu, Tingchen Fu, Xinting Huang, Enbo Zhao, Yu Zhang, Yulong Chen, Longyue Wang, Anh Tuan Luu, Wei Bi, Freda Shi, Shuming Shi

Computational Linguistics

Code / Paper

A survey of hallucination in LLMs.

Multilinguality

Lost in Literalism: How Supervised Training Shapes Translationese in LLMs

Yafu Li^*, Ronghao Zhang^*, Zhilin Wang, Huajiang Zhang, Leyang Cui, Yongjing Yin, Tong Xiao, Yue Zhang

ACL, 2025

Code / Paper

A systematic study of the origin of translationese in LLMs and mitigation methods.

Explicit Syntactic Guidance for Neural Text Generation

Yafu Li, Leyang Cui, Jianhao Yan, Yongjing Yin, Wei Bi, Shuming Shi, Yue Zhang

ACL, 2023; Best Paper Nomination

Code / Paper

A neural-symbolic method that guides generation with rules.

Multi-Granularity Optimization for Non-Autoregressive Translation

Yafu Li, Leyang Cui, Yongjing Yin, Yue Zhang

EMNLP, 2022

Code / Paper

Optimizing non-autoregressive translation with multi-granularity policy gradient.

On Compositional Generalization of Neural Machine Translation

Yafu Li, Yongjing Yin, Yulong Chen, Yue Zhang

ACL, 2021; Oral

Code / Paper

Neural machine translation suffers poor compositionality.

Mentees

I closely supervise and mentor research interns and students. Selected mentees and representative publications include:

• Runzhe Zhan: ICLR 2026
• Zican Hu: ICLR 2026
• Shilin Zhang: ICLR 2026
• Guanxu Chen: ICLR 2026
• Yuxian Jiang: preprint
• Haoran Zhang: preprint
• Jianhao Yan: NeurIPS 2025
• Tingchen Fu: ACL 2026
• Sen Yang: ACL 2026 Findings
• Ronghao Zhang: ACL 2025
• Zhilin Wang: ACL 2024 Findings; ACL 2025; ACL 2025 Findings; ACL 2026 Findings
• Huajian Zhang: ACL 2024 Findings

We are looking for full-time researchers, research interns, and joint PhD students with THU, PKU, SJTU, FDU, and other universities to work on cutting-edge research in large language models.

Talks & Lectures

• October 27, 2025 - Southern University of Science and Technology: On the Evolution of Reasoning Abilities in Large Language Models
• September 4, 2025 - Tencent: Evolving Reasoning Abilities of LLMs: RLVR, Off-Policy Learning, and Test-Time Reinforcement Learning
• August 12, 2025 - CCL 2025 Forum on Large Model Reasoning and Reinforcement Learning: Invited speaker

Service

Honors

• Outstanding Student Scholarship, Silver medal, Tencent Rhino-Bird Elite Program, 2024.
• National Scholarship, Ministry of Education, 2023.
• Dean's Medal, Westlake University, 2023.