Shengnan An

Shengnan An

Hi! I’m Shengnan An (安晟男), a researcher on the LongCat Foundation Model Team at Meituan, which I joined in July 2025 through the Beidou graduate program.

I received my Ph.D. from Xi’an Jiaotong University in June 2025, advised by Prof. Nanning Zheng, through a joint training program with Microsoft Research Asia (MSRA). I was a research intern at MSRA from August 2019 to June 2025, working with Jian-Guang Lou and Zeqi Lin.

My research focuses on understanding and improving the capabilities of large language models, with an emphasis on mid-training data curation, agents for productivity work, and mathematical and general reasoning. My earlier research covered long-context understanding, in-context learning, code generation, and compositional generalization.

Google Scholar · Semantic Scholar · Email

💼 Experience

Meituan · LongCat Foundation Model Team

July 2025 – Present · Researcher, Beidou Program

I contribute to the development and releases of the LongCat model family, focusing on mid-training data curation and model evaluation.

  • Mid-training data curation. Turn user interaction logs into quality-controlled agent trajectories through preprocessing, session aggregation, filtering, and de-identification. Analyze data retention across sources and batches, and use small-model training experiments to inform data admission.
  • AI at Work evaluation. Build and iterate on tasks, environments, and rubric-based LLM grading for productivity scenarios. Analyze evaluation failures and adapt the evaluation pipeline for comparisons with external models. [FORTE]
  • Reasoning evaluation. Build challenging benchmarks for mathematical and general reasoning, combining reliable automated grading with analyses of token efficiency, test-time scaling, and cross-lingual robustness. [AMO-Bench] [General365]

Microsoft Research Asia · Research Intern

August 2019 – June 2025

  • Long-context understanding. Developed information-intensive training and FILM-7B to address information loss in long contexts. The training data was adopted in the Microsoft Phi-3 model family. [FILM]
  • Learning from mistakes. Explored mistake-correction training to improve mathematical reasoning and data efficiency. [LEMA]
  • In-context learning and code generation. Developed skill-based example selection to improve semantic parsing and code generation, with related techniques adopted in Microsoft DBCopilot.
  • Compositional generalization. Systematically evaluated language models’ ability to recombine learned concepts, and developed lightweight models with near-human compositional generalization performance.

📝 Publications

  • General365: Benchmarking General Reasoning in Large Language Models Across Diverse and Challenging Tasks. Preprint, 2026. [Paper] [Project]
    Junlin Liu, Shengnan An, Shuang Zhou, Dan Ma, Shixiong Luo, Ying Xie, Yuan Zhang, Wenling Yuan, Yifan Zhou, Xiaoyu Li, Ziwen Wang, Xuezhi Cao, Xunliang Cai.

  • AMO-Bench: Large Language Models Still Struggle in High School Math Competitions. Findings of ACL 2026. [Paper] [Project]
    Junlin Liu, Shengnan An, Shuang Zhou, Dan Ma, Yehao Lin, Xinxuan Lv, Xuanlin Wang, Xiaoyu Li, Ziwen Wang, Xuezhi Cao, Xunliang Cai.

  • WorkThrough-Bench: Benchmarking Agents on Long-Horizon Productivity Work.
    Xin Gao, Jiamu Li, Yuhuai Wei, Zhifeng Li, Weikang Zhang, Shengnan An, Hansi Yang, Hongyan Hao, Zijian Zhang, Mingyang Zhu, Shengyu Guo, Pu Miao, Xuhang Zhu, Linsen Guo, Ziwen Wang, Xiaoyu Li, Xuezhi Cao.

  • Make Your LLM Fully Utilize the Context. NeurIPS 2024. [Paper] [Code]
    Shengnan An, Zexiong Ma, Zeqi Lin, Nanning Zheng, Jian-Guang Lou, Weizhu Chen.

  • Can LLMs Learn From Mistakes? An Empirical Study on Reasoning Tasks. Findings of EMNLP 2024. [Paper] [Code]
    Shengnan An, Zexiong Ma, Siqi Cai, Zeqi Lin, Nanning Zheng, Jian-Guang Lou, Weizhu Chen.

  • Learning From Mistakes Makes LLM Better Reasoner. Preprint version of the EMNLP 2024 work. [Paper] [Code]
    Shengnan An, Zexiong Ma, Zeqi Lin, Nanning Zheng, Jian-Guang Lou, Weizhu Chen.

  • Compositional API Recommendation for Library-Oriented Code Generation. ICPC 2024. [Paper]
    Zexiong Ma, Shengnan An, Bing Xie, Zeqi Lin.

  • Skill-Based Few-Shot Selection for In-Context Learning. EMNLP 2023. [Paper]
    Shengnan An, Bo Zhou, Zeqi Lin, Qiang Fu, B. Chen, Nanning Zheng, Weizhu Chen, Jian-Guang Lou.

  • How Do In-Context Examples Affect Compositional Generalization? ACL 2023. [Paper] [Code]
    Shengnan An, Zeqi Lin, Qiang Fu, Bei Chen, Nanning Zheng, Jian-Guang Lou, Dongmei Zhang.

  • Does Deep Learning Learn to Abstract? A Systematic Probing Framework. ICLR 2023. [Paper] [Code]
    Shengnan An, Zeqi Lin, Bei Chen, Qiang Fu, Nanning Zheng, Jian-Guang Lou.

  • Learning Algebraic Recombination for Compositional Generalization. Findings of ACL-IJCNLP 2021. [Paper] [Code]
    Chenyao Liu*, Shengnan An*, Zeqi Lin, Qian Liu, Bei Chen, Jian-Guang Lou, Lijie Wen, Nanning Zheng, Dongmei Zhang.

  • Compositional Generalization by Learning Analytical Expressions. NeurIPS 2020 · Spotlight. [Paper] [Code]
    Qian Liu*, Shengnan An*, Jian-Guang Lou, Bei Chen, Zeqi Lin, Yan Gao, Bin Zhou, Nanning Zheng, Dongmei Zhang.

  • Enhancing Retrieval-Augmented Generation with Dehallucinating Parallel Context Extension. ICML 2025 Workshop on Long-Context Foundation Models. [Paper]
    Zexiong Ma, Shengnan An, Zeqi Lin, Yanzhen Zou, Jian-Guang Lou, Bing Xie.

  • Dehallucinating Parallel Context Extension for Retrieval-Augmented Generation. Preprint, 2024. [Paper]
    Zexiong Ma, Shengnan An, Zeqi Lin, Yanzhen Zou, Jian-Guang Lou, Bing Xie.

  • Input-Tuning: Adapting Unfamiliar Inputs to Frozen Pretrained Models. Preprint, 2022. [Paper]
    Shengnan An*, Yifei Li*, Zeqi Lin, Qian Liu, Bei Chen, Qiang Fu, Weizhu Chen, Nanning Zheng, Jian-Guang Lou.

  • STAND-Guard: A Small Task-Adaptive Content Moderation Model. COLING 2025. [Paper]
    Minjia Wang, Pingping Lin, Siqi Cai, Shengnan An, Shengjie Ma, Zeqi Lin, Congrui Huang, Bixiong Xu.

* Equal contribution.

📖 Education

  • Xi’an Jiaotong University, Ph.D., September 2020 – June 2025. Advisor: Prof. Nanning Zheng. Joint training with Microsoft Research Asia; research in natural language processing.
  • Xi’an Jiaotong University, B.Eng. in Automation, September 2016 – June 2020. Qian Xuesen Honors Class; GPA: 4.0/4.3.

☀️ Honors and Awards

  • Outstanding Graduate (2025)
  • National Scholarship (2024)
  • China Mobile Scholarship (2024)
  • Outstanding Graduate Student (2023)
  • Outstanding Graduate Cadre (2020)
  • Star of Tomorrow in MSRA (2019, 2020)
  • Outstanding Student Cadre (2017, 2018, 2019)
  • First Class Scholarship (2017, 2019)

🌿 Academic Service

Reviewer for: NeurIPS, ICLR, ICML, ACL Rolling Review（ACL, EMNLP, NAACL, EACL）