Yifan Hao
Research

Algorithms, with foundations.

My work connects practical questions in LLM post-training and reinforcement learning with statistical learning theory, generalization, and robustness.

Publications & preprints

  1. Preprint

    Muon for RLVR: What Works, Why, and What It Learns

    Yifan Hao, Zoey (Sha) Li, Changlong Yu, Qin Lu, Shuowei Jin, Xin Liu, Zhenyu Shi, Tong Zhang

  2. NewInML Workshop @ NeurIPS2026

    Retrieval-Guided Fine-Tuning as Noisy Estimation: Risk bounds and Architectural Analysis

    Bhargav Lad, Yifan Hao

    Paper
  3. arXiv

    Adaptive Layerwise Perturbation: Unifying Off-Policy Corrections for LLM RL

    Chenlu Ye*, Xuanchang Zhang*, Yifan Hao*, Zhou Yu, Ziji Zhang, Abhinav Gullapalli, Hao Chen, Jing Huang, Tong Zhang

    Paper
  4. TMLR2026

    Task Vector Bases: A Unified and Scalable Framework for Compressed Task Arithmetic

    Siqi Zeng, Yifei He, Meitong Liu, Weiqiu You, Yifan Hao, Yao-Hung Hubert Tsai, Makoto Yamada, Han Zhao

    Paper Code
  5. IEEE Transactions on Information Theory

    The Surprising Harmfulness of Benign Overfitting for Adversarial Robustness

    Yifan Hao, Tong Zhang

    Paper Journal
  6. Blog · In progress

    Revisiting the Predictability of RLVR

    Blog
  7. arXiv

    Towards Better Generalization via Distributional Input Projection Network

    Yifan Hao*, Yanxin Lu*, Hanning Zhang, Xinwei Shen, Tong Zhang

    Paper
  8. arXiv

    Transformers as Multi-task Learners: Decoupling Features in Hidden Markov Models

    Yifan Hao*, Chenlu Ye*, Chi Han, Tong Zhang

    Paper
  9. NeurIPS 2025

    Optimizing Chain-of-Thought Reasoners via Gradient Variance Minimization in Rejection Sampling and RL

    Jiarui Yao*, Yifan Hao*, Hanning Zhang, Hanze Dong, Wei Xiong, Nan Jiang, Tong Zhang

    Paper
  10. ICML 2025

    Understanding Overadaptation in Supervised Fine-Tuning: The Role of Ensemble Methods

    Yifan Hao*, Xingyuan Pan*, Hanning Zhang*, Chenlu Ye, Rui Pan, Tong Zhang

    Paper
  11. Journal of Multivariate Analysis

    SFQRA: Scaled Factor-augmented Quantile Regression with Aggregation in Conditional Mean Forecasting

    Lei Shu, Yifan Hao, Yu Chen, Qing Yang

    Paper
  12. arXiv

    On the Benefits of Over-parameterization for Out-of-Distribution Generalization

    Yifan Hao*, Yong Lin*, Difan Zou, Tong Zhang

    Paper
  13. ICLR 2024

    Spurious feature diversification improves out-of-distribution generalization

    Yong Lin*, Lu Tan*, Yifan Hao*, Honam Wong, Hanze Dong, Weizhong Zhang, Yujiu Yang, Tong Zhang

    Paper
  14. ICLR 2024

    Reverse Diffusion Monte Carlo

    Xunpeng Huang*, Hanze Dong*, Yifan Hao, Yi-An Ma, Tong Zhang

    Paper

* indicates alphabetical author order or equal contribution.

Let’s talk research.

I welcome conversations about LLM post-training, reinforcement learning, AI agents, and full-time or internship opportunities.

yifanh12@illinois.edu