Publications

2026

1. COLM

   Token-Level Off-Policy Learning for Faithful Generation Under Distribution Shift

   Zitong Huang^*, Gustavo Lucas Carvalho^*, Deqing Fu, and Robin Jia

   In Conference on Language Modeling (COLM), 2026

   📄 Paper
2. ICML

   

   Transformers Provably Learn Algorithmic Solutions for Graph Connectivity, But Only with the Right Data

   Qilin Ye^*, Deqing Fu^*, Robin Jia, and Vatsal Sharan

   In International Conference on Machine Learning (ICML), 2026

   *Equal Contribution

   📄 Paper PDF
3. arXiv

   Value-Aware Stochastic KV Cache Eviction for Reasoning Models

   Ting-Yun Chang^* , Harvey Yiyun Fu^*, Deqing Fu^*, Chenghao Yang, Jesse Thomason, and Robin Jia

   In arXiv, 2026

   *Equal Contribution

   📄 Paper
4. COLM

   

   Convergent Evolution: How Different Language Models Learn Similar Number Representations

   Deqing Fu, Tianyi Zhou, Mikhail Belkin, Vatsal Sharan, and Robin Jia

   In Conference on Language Modeling (COLM), 2026

   📄 Paper Blog Website
5. ICML

   

   EPSVec: Efficient and Private Synthetic Data Generation via Dataset Vectors

   Amin Banayeeanzade, Qingchuan Yang, Deqing Fu, Spencer Hong, Erin Babinsky, Alfy Samuel, Anoop Kumar, Robin Jia, and Sai Praneeth Karimireddy

   In International Conference on Machine Learning (ICML), 2026

   📄 Paper
6. arXiv

   

   Are LLM Decisions Faithful to Verbal Confidence?

   Jiawei Wang , Yanfei Zhou, Siddartha Devic, and Deqing Fu

   2026

   📄 Paper
7. ICLR

   

   Zebra-CoT: A Dataset for Interleaved Vision Language Reasoning

   Ang Li^*, Charles Wang^*, Deqing Fu^*, Kaiyu Yue^*, Zikui Cai^*, Wang Bill Zhu^*, Ollie Liu^* , Peng Guo^*, Willie Neiswanger, Furong Huang, Tom Goldstein, and Micah Goldblum

   In International Conference on Learning Representations (ICLR), 2026

   *Equal Contribution

   📄 Paper Dataset
8. ICLR

   

   FoNE: Precise Single-Token Number Embeddings via Fourier Features

   Tianyi Zhou, Deqing Fu, Mahdi Soltanolkotabi, Robin Jia, and Vatsal Sharan

   In International Conference on Learning Representations (ICLR), 2026

   📄 Paper Website
9. COLM

   

   Resa: Transparent Reasoning Models via SAEs

   Shangshang Wang, Julian Asilis, Ömer Faruk Akgül, Enes Burak Bilgin, Ollie Liu, Deqing Fu, and Willie Neiswanger

   In Conference on Language Modeling (COLM), 2026

   📄 Paper
10. ACL

    

    Textual Steering Vectors Can Improve Visual Understanding in Multimodal Large Language Models

    Woody Haosheng Gan^*, Deqing Fu^*, Julian Asilis^*, Ollie Liu^*, Dani Yogatama, Vatsal Sharan, Robin Jia, and Willie Neiswanger

    In Association of Computational Linguistics (ACL), 2026

    *Equal Contribution

    📄 Paper Blog

2025

1. Tech Report

   Technical Report: Qwen3 Cross-layer Transcoders

   Jakob Hansen, Deqing Fu , Shangshang Wang, and John Carlsson

   2025

   📄 Paper
2. NeurIPS

   

   VisualLens: Personalization through Visual History

   Wang Bill Zhu, Deqing Fu , Kai Sun, Yi Lu, Zhaojiang Lin, Seungwhan Moon, Kanika Narang, Mustafa Canim , Yue Liu, Anuj Kumar, and Xin Luna Dong

   In Conference on Neural Information Processing Systems (NeurIPS), 2025

   📄 Paper
3. ICLR

   

   TLDR: Token-Level Detective Reward Model for Large Vision Language Models

   Deqing Fu, Tong Xiao , Rui Wang, Wang Zhu, Pengchuan Zhang, Guan Pang, Robin Jia, and Lawrence Chen

   In International Conference on Learning Representations (ICLR), 2025

   📄 Paper
4. ICLR

   

   Transformers Learn Low Sensitivity Functions: Investigations and Implications

   Bhavya Vasudeva^*, Deqing Fu^*, Tianyi Zhou, Elliot Kau , You-Qi Huang, and Vatsal Sharan

   In International Conference on Learning Representations (ICLR), 2025

   *Equal Contribution

   📄 Paper
5. ICLR

   

   DeLLMa: Decision Making Under Uncertainty with Large Language Models

   Ollie Liu^*, Deqing Fu^*, Dani Yogatama, and Willie Neiswanger

   In International Conference on Learning Representations (ICLR), 2025

   Spotlight (Top 5.1%), *Equal Contribution

   📄 Paper Code Website
6. NAACL

   

   DreamSync: Aligning Text-to-Image Generation with Image Understanding Feedback

   Jiao Sun^*, Deqing Fu^*, Yushi Hu^* , Su Wang, Royi Rassin, Da-Cheng Juan, Dana Alon, Charles Herrmann, Sjoerd Steenkiste, Ranjay Krishna, and Cyrus Rashtchian

   In Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL), 2025

   *Equal Contribution

   📄 Paper

2024

1. NeurIPS

   

   Transformers Learn to Achieve Second-Order Convergence Rates for In-Context Linear Regression

   Deqing Fu, Tian-Qi Chen, Robin Jia, and Vatsal Sharan

   In Conference on Neural Information Processing Systems (NeurIPS), 2024

   SoCalNLP Symposium 2023 Best Paper Award

   📄 Paper Code
2. NeurIPS

   

   Pre-trained Large Language Models Use Fourier Features to Compute Addition

   Tianyi Zhou, Deqing Fu, Vatsal Sharan, and Robin Jia

   In Conference on Neural Information Processing Systems (NeurIPS), 2024

   📄 Paper
3. COLM

   

   IsoBench: Benchmarking Multimodal Foundation Models on Isomorphic Representations

   Deqing Fu^*, Ruohao Guo^*, Ghazal Khalighinejad^*, Ollie Liu^*, Bhuwan Dhingra, Dani Yogatama, Robin Jia, and Willie Neiswanger

   In Conference on Language Modeling (COLM), 2024

   *Equal Contribution

   📄 Paper Website

2023

1. EMNLP

   

   SCENE: Self-Labeled Counterfactuals for Extrapolating to Negative Examples

   Deqing Fu, Ameya Godbole, and Robin Jia

   In Conference on Empirical Methods in Natural Language Processing (EMNLP), 2023

   📄 Paper Code