Zinan Tang (εζ’ζ₯ in Chinese) is currently pursuing his masterβs degree in Electronic and Information Engineering (EIE) at the Department of Automation (DA), Tsinghua University (THU), under the supervision of Dr. Biqing Huang. He received his B.S. degree in Computer Science (CS) from the School of Computer Science (Pilot Software Engineering School, SCS) at Beijing University of Posts and Telecommunication (BUPT). He currently is a Research Intern (RI) of Ant Group, participating in training data optimization for the Ant Ling AI Model Family.
His research interests are on Data-centric LLMs (e.g., AI4data, data4AI, SFT, post-training, pre-training, reasoning).
π₯ News
- 2026.07: π₯π₯ CausalMix is released on arXiv!
- 2026.04: ππ REST is accepted by ACL 2026 (Main). Congrats to Zhuoshi Pan!
- 2025.12: π₯π₯ The technical report of OpenDataArena is on arXiv!
- 2025.09: π₯π₯ ScaleDiff is released on arXiv!
- 2025.08: ππ Middo is accepted by EMNLP 2025 (Main). Thanks for all collaborators!
π Publications
β (Co) First-authored Publications
CausalMix: Data Mixture as Causal Inference for Language Model Training
Zinan Tang, Yukun Zhang, Shaomian Zheng, Zhuoshi Pan, Qizhi Pei, Dingnan Jin, Jun Zhou, Yujun Wang, Biqing Huang
Middo: Model-Informed Dynamic Data Optimization for Enhanced LLM Fine-Tuning via Closed-Loop Learning
Zinan Tang, Xin Gao, Zhuoshi Pan, Qizhi Pei, Mengzhang Cai, Jiang Wu, Conghui He, Lijun Wu
Project | 
MTRBench: A Multimodal Reasoning Benchmark from Reality Shows (fka. Big Escape Benchmark: Evaluating Human-Like Reasoning in Language Models via Real-World Escape Room Challenges)
Zinan Tang, QiYao Sun, Zhuoshi Pan, Qizhi Pei, Xin Gao, Mengyuan Sun, Honglin Lin, Mengzhang Cai, Yu Li, Chenlin Ming, Jiang Wu, Conghui He, Lijun Wu
π° Technical Reports
OpenDataArena: A Fair and Open Arena for Benchmarking Post-Training Dataset Value
Mengzhang Cai, Xin Gao, Yu Li, Honglin Lin, Zheng Liu, Zhuoshi Pan, Qizhi Pei, Xiaoran Shang, Mengyuan Sun, Zinan Tang, Xiaoyang Wang, Zhanping Zhong, Yun Zhu, Dahua Lin, Conghui He, Lijun Wu
Project | 
π€ Co-authored Publications
REST: Stress Testing Large Reasoning Models by Asking Multiple Problems at Once
Zhuoshi Pan, Qizhi Pei, Yu Li, Zinan Tang, Qiyao Sun, H. Vicky Zhao, Conghui He, Lijun Wu
Project |
ScaleDiff: Scaling Difficult Problems for Advanced Mathematical Reasoning
Qizhi Pei, Zhuoshi Pan, Honglin Lin, Xin Gao, Yu Li, Zinan Tang, Conghui He, Rui Yan, Lijun Wu
Project | 
Davidβs Slingshot: A Strategic Coordination Framework of Small LLMs Matches Large LLMs in Data Synthesis
Xin Gao, Qizhi Pei, Zinan Tang, Yu Li, Honglin Lin, Jiang Wu, Lijun Wu, Conghui He
Project | 
LEMMA: Learning from Errors for MatheMatical Advancement in LLMs
Zhuoshi Pan, Yu Li, Honglin Lin, Qizhi Pei, Zinan Tang, Wei Wu, Chenlin Ming, H. Vicky Zhao, Conghui He, Lijun Wu
Project | 
π Honors and Awards
- 2025, National Scholarship, Ministry of Education, PRC
- 2024, Second Prize Scholarship, BUPT
- 2023, National Scholarship, Ministry of Education, PRC
π Educations
- 2026.09 - 2028.06 (Expected), masterβs student in DA, THU, major in EIE.
- 2022.09 - 2026.06, undergraduate student in SCS, BUPT, major in CS.
π» Internships
- 2025.12 - present, Ant Group, Ant Ling AI Model Family, Data Optimization, Beijing, China
- 2024.09 - 2025.10, Shanghai Artificial Intelligent Laboratory, OpenDataLab, RAISE, Beijing, China
- 2024.07 - 2024.08, University of Science and Technology of China (USTC), State Key Laboratory of Cognitive Intelligence, COGAI, Hefei, Anhui, China.
- 2023.07 - 2024.06, BUPT, Department of Intelligent Science and Technology, Beijing, China.
π¬ Academic Service
- Reviewer: NeurIPS 2026