Publications

Ziyang Cai*, Xingyu Zhu*, Yihe Dong, Yinghui He, Sanjeev Arora (2026). T2MLR: Transformer with Temporal Middle-Layer Recurrence. LIT Workshop @ ICLR 2026.

Yun Cheng, Xingyu Zhu, Haoyu Zhao, Sanjeev Arora (2026). Contextual Drag - How Errors in the Context Affect LLM Reasoning. arXiv.

Xingyu Zhu*, Abhishek Panigrahi*, Sanjeev Arora (2025). On the Power of Context-Enhanced Learning in LLMs. ICML 2025 (Spotlight).

Xingyu Zhu*, Zixuan Wang*, Xiang Wang, Mo Zhou, Rong Ge (2023). Understanding Edge-of-Stability Training Dynamics with a Minimalist Example. ICLR 2023.

Zeyu Shen, Zhiyi Wang, Xingyu Zhu, Brandon Fain, Kamesh Munagala (2023). Fairness in the Assignment Problem with Uncertain Priorities. AAMAS 2023.

Yikai Wu*, Xingyu Zhu*, Chenwei Wu, Annie Wang, Rong Ge (2022). Dissecting Hessian: Understanding Common Structure of Hessian in Neural Networks. arXiv.