Tian Liang

Researcher, Tencent AI Lab

Joined

February 2022

Names

Tian Liang (Preferred)

Emails

****@mails.tsinghua.edu.cn (Confirmed)

****@tencent.com (Confirmed)

****@gmail.com (Confirmed)

Personal Links

Career & Education History

Researcher

Tencent AI Lab (tencent.com)

2024 – Present

MS student

Graduate School at Shenzhen,Tsinghua University (tsinghua.edu.cn)

2021 – 2024

Advisors, Relations & Conflicts

Coauthor

Haitao Mi

2024 – Present

Coauthor

Dong Yu

2024 – Present

Coauthor

Zicheng Lin

2022 – Present

Coauthor

Zhaopeng Tu

2021 – Present

Coauthor

Xing Wang

2021 – Present

Coauthor

Shuming Shi

2021 – Present

Coauthor

Wenxiang Jiao

2021 – Present

Coauthor

Zhiwei He

2021 – Present

Coauthor

Yujiu Yang

2021 – Present

Coauthor

Rui Wang

2021 – Present

Expertise

No areas of expertise listed

Publications

DeepCompress: A Dual Reward Strategy for Dynamically Exploring and Compressing Reasoning Chains
Tian Liang, Wenxiang Jiao, Zhiwei He, Jiahao Xu, Haitao Mi, Dong Yu
- ICLR 2026 Poster
- Readers: Everyone
DeepTheorem: Advancing LLM Reasoning for Theorem Proving Through Natural Language and Reinforcement Learning
Ziyin Zhang, Jiahao Xu, Zhiwei He, Tian Liang, Qiuzhi Liu, Yansi Li, Linfeng Song, Zhenwen Liang, Zhuosheng Zhang, Rui Wang, Zhaopeng Tu, Haitao Mi, Dong Yu
- Submitted to ICLR 2026
- Readers: Everyone
DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoning
Zhiwei He, Tian Liang, Jiahao Xu, Qiuzhi Liu, Xingyu Chen, Yue Wang, Linfeng Song, Dian Yu, Zhenwen Liang, Wenxuan Wang, Zhuosheng Zhang, Rui Wang, Zhaopeng Tu, Haitao Mi, Dong Yu
- ICLR 2026 Poster
- Readers: Everyone
Dynamic Parametric Retrieval Augmented Generation
Yuqiao Tan, Shizhu He, Huanxuan Liao, Tian Liang, Jun Zhao, Kang Liu
- ICLR 2026 Conference Withdrawn Submission
- Readers: Everyone
The Pensieve Paradigm: Stateful Language Models Mastering Their Own Context
Xiaoyuan Liu, Tian Liang, Dongyang Ma, Deyu Zhou, Haitao Mi, Pinjia He, Yan Wang
- ICLR 2026 Poster
- Readers: Everyone
The First Few Tokens Are All You Need: An Efficient and Effective Unsupervised Prefix Fine-Tuning Method for Reasoning Models
Ke Ji, Jiahao Xu, Tian Liang, Qiuzhi Liu, Zhiwei He, Xiaoyuan Liu, Xingyu Chen, Junying Chen, Benyou Wang, Zhaopeng Tu, Haitao Mi, Dong Yu
- NeurIPS 2025 poster
- Readers: Everyone
Two Experts Are All You Need for Steering Thinking: Reinforcing Cognitive Effort in MoE Reasoning Models Without Additional Training
Mengru Wang, Xingyu Chen, Yue Wang, Zhiwei He, Jiahao Xu, Tian Liang, Qiuzhi Liu, Yunzhi Yao, Wenxuan Wang, Ruotian Ma, Haitao Mi, Ningyu Zhang, Zhaopeng Tu, Xiaolong Li, Dong Yu
- NeurIPS 2025 poster
- Readers: Everyone
Thoughts Are All Over the Place: On the Underthinking of Long Reasoning Models
Yue Wang, Qiuzhi Liu, Jiahao Xu, Tian Liang, Xingyu Chen, Zhiwei He, Linfeng Song, Dian Yu, Juntao Li, Zhuosheng Zhang, Rui Wang, Zhaopeng Tu, Haitao Mi, Dong Yu
- NeurIPS 2025 spotlight
- Readers: Everyone
Trust, But Verify: A Self-Verification Approach to Reinforcement Learning with Verifiable Rewards
Xiaoyuan Liu, Tian Liang, Zhiwei He, Jiahao Xu, Wenxuan Wang, Pinjia He, Zhaopeng Tu, Haitao Mi, Dong Yu
- NeurIPS 2025 poster
- Readers: Everyone
Dancing with Critiques: Enhancing LLM Reasoning with Stepwise Natural Language Self-Critique
Yansi Li, Jiahao Xu, Tian Liang, Xingyu Chen, Zhiwei He, Qiuzhi Liu, Rui Wang, Zhuosheng Zhang, Zhaopeng Tu, Haitao Mi, Dong Yu
- CoRR 2025
- Readers: Everyone

View all 39 publications

Co-Authors

View all 60 co-authors