Zefan Cai, whose chinese name is 蔡泽凡, is a Ph.D student in University of Wisconsin - Madison. During his master, he did brode research about LLM topics. In his Ph.D, He mainly did research on sparse attention, KV cache compression and other inference optimization topics. He considers long-context as the most important problem in LLM.

Study

• Ph.D: University of Wisconsin - Madison
  • Advisor: Junjie Hu
  • September, 2024 - June, 2028 (Expected)
• M.S.: Peking University
  • Advisor: Baobao Chang
  • September, 2022 - June, 2024
• B.S.: Beijing Jiaotong University
  • September, 2018 - June, 2022

Intern

• Adobe Research - San Jose, California
  • Advisor: Jiuxiang Gu and Hao Tan
  • May 2025 - August 2025 (Expected)
• Microsoft Azure
  • Advisor: Wen Xiao
• Qwen - Beijing
  • Advisor: Tianyu Liu and Keming Lu

News

• June 2025 I join Adobe Research as a research intern!
• May 2025 One paper (Adaptivestep) accepted by ICML 2025!
• May 2025 One paper (Math-Minos) accepted by ACL 2025!
• Jan. 2025 Four papers (DnD-Transformer, HeadKV, Omni-MATH, GDPO) accepted by ICLR 2025!
• August 2024 One paper (VeCAF) accepted by ACM Multimedia 2024!
• May 2025 Four papers (PCA-Bench, ZeroED, CENSOR, FairEval) accepted by ACL 2025!
• Mar. 2024 Two papers (DialogVCS, ALSACE) accepted by NAACL 2024!
• Jan. 2024 One paper (MMICL) accepted by ICLR 2025!
• May 2023 One paper (SANTA) accepted by ACL 2023!
• Jan. 2023 One paper (CTR) accepted by ICLR 2023!

Selected Projects

R-KV: Redundancy-aware KV Cache Compression for Training-Free Reasoning Models Acceleration

Zefan Cai, Wen Xiao, Hanshi Sun, Cheng Luo, Yikai Zhang, Ke Wan, Yucheng Li, Yeyang Zhou, Li-Wen Chang, Jiuxiang Gu, Zhen Dong, Anima Anandkumar, Abedelkadir Asi, Junjie Hu
Preprint.
PDF Page Code Twitter Youtube

PyramidKV: Dynamic KV Cache Compression based on Pyramidal Information Funneling

Zefan Cai, Yichi Zhang, Bofei Gao, Yuliang Liu, Yucheng Li, Tianyu Liu, Keming Lu, Wayne Xiong, Yue Dong, Junjie Hu, Wen Xiao
Preprint.
PDF Page Code Twitter Article 1 Article 2 Article 3 Article 4

All Projects

R-KV: Redundancy-aware KV Cache Compression for Training-Free Reasoning Models Acceleration

Zefan Cai, Wen Xiao, Hanshi Sun, Cheng Luo, Yikai Zhang, Ke Wan, Yucheng Li, Yeyang Zhou, Li-Wen Chang, Jiuxiang Gu, Zhen Dong, Anima Anandkumar, Abedelkadir Asi, Junjie Hu
Preprint.
PDF Page Code Twitter

From Preferences to Prejudice: The Role of Alignment Tuning in Shaping Social Bias in Video Diffusion Models

Zefan Cai*, Haoyi Qiu*, Haozhe Zhao*, Ke Wan, Jiachen Li, Jiuxiang Gu, Wen Xiao, Nanyun Peng, Junjie Hu
Preprint.
PDF

MENTOR: Efficient Multimodal-Conditioned Tuning for Autoregressive Vision Generation Models

Haozhe Zhao*, Zefan Cai*, Shuzheng Si, Liang Chen, Jiuxiang Gu, Wen Xiao, Junjie Hu
Preprint.
PDF

VisualToolAgent (VisTA): A Reinforcement Learning Framework for Visual Tool Selection

Zeyi Huang, Yuyang Ji, Anirudh Sundara Rajan, Zefan Cai, Wen Xiao, Junjie Hu, Yong Jae Lee
Preprint.
PDF Page Code Twitter 1 Twitter 2 Twitter 3

Adaptivestep: Automatically dividing reasoning step through model confidence

Yuliang Liu, Junjie Lu, Zhaoling Chen, Chaofeng Qu, Jason Klein Liu, Chonghan Liu, Zefan Cai, Yunhui Xia, Li Zhao, Jiang Bian, Chuheng Zhang, Wei Shen, Zhouhan Lin
ICML 2025 (Poster)
PDF Code Model Data

Headinfer: Memory-efficient llm inference by head-wise offloading

Cheng Luo, Zefan Cai, Hanshi Sun, Jinqi Xiao, Bo Yuan, Wen Xiao, Junjie Hu, Jiawei Zhao, Beidi Chen, Anima Anandkumar
ICML 2025 Workshop on Long Context Foundation Models (LCFM)
PDF Code

No Preference Left Behind: Group Distributional Preference Optimization

Binwei Yao, Zefan Cai, Yun-Shiuan Chuang, Shanglin Yang, Ming Jiang, Diyi Yang, Junjie Hu
ICLR 2025 (Poster)
PDF Code

Next Token Prediction Towards Multimodal Intelligence: A Comprehensive Survey

Liang Chen, Zekun Wang, Shuhuai Ren, Lei Li, Haozhe Zhao, Yunshui Li, Zefan Cai, Hongcheng Guo, Lei Zhang, Yizhe Xiong, Yichi Zhang, Ruoyu Wu, Qingxiu Dong, Ge Zhang, Jian Yang, Lingwei Meng, Shujie Hu, Yulong Chen, Junyang Lin, Shuai Bai, Andreas Vlachos, Xu Tan, Minjia Zhang, Wen Xiao, Aaron Yee, Tianyu Liu, Baobao Chang
Priprint.
PDF Repo

Not all heads matter: A head-level kv cache compression method with integrated retrieval and reasoning

Yu Fu, Zefan Cai, Abedelkadir Asi, Wayne Xiong, Yue Dong, Wen Xiao
ICLR 2025 (Poster)
PDF Code

A spark of vision-language intelligence: 2-dimensional autoregressive transformer for efficient finegrained image generation

Liang Chen, Sinan Tan, Zefan Cai, Weichu Xie, Haozhe Zhao, Yichi Zhang, Junyang Lin, Jinze Bai, Tianyu Liu, Baobao Chang
ICLR 2025 (Poster)
PDF Code Model

COMMA: A Communicative Multimodal Multi-Agent Benchmark

Timothy Ossowski, Jixuan Chen, Danyal Maqbool, Zefan Cai, Tyler Bradshaw, Junjie Hu
Preprint.
PDF

Omni-math: A universal olympiad level mathematic benchmark for large language models

Bofei Gao, Feifan Song, Zhe Yang, Zefan Cai, Yibo Miao, Qingxiu Dong, Lei Li, Chenghao Ma, Liang Chen, Runxin Xu, Zhengyang Tang, Benyou Wang, Daoguang Zan, Shanghaoran Quan, Ge Zhang, Lei Sha, Yichang Zhang, Xuancheng Ren, Tianyu Liu, Baobao Chang
ICLR 2025 (Poster)
PDF Page Repo Code Data Model

ML-bench: Large language models leverage open-source libraries for machine learning tasks

Yuliang Liu*, Xiangru Tang*, Zefan Cai*, Junjie Lu, Yichi Zhang, Yanjun Shao, Zexuan Deng, Helan Hu, Zengxian Yang, Kaikai An, Ruijun Huang, Shuzheng Si, Sheng Chen, Haozhe Zhao, Zhengliang Li, Liang Chen, Yiming Zong, Yan Wang, Tianyu Liu, Zhiwei Jiang, Baobao Chang, Yujia Qin, Wangchunshu Zhou, Yilun Zhao, Arman Cohan, Mark Gerstein
ICLR 2025 Deep Learning for Code (DL4C) Workshop
ICLR 2025 Agentic AI for Scientific Discovery Workshop
PDF Page Code Data

Towards a unified view of preference learning for large language models: A survey

Bofei Gao, Feifan Song, Yibo Miao, Zefan Cai, Zhe Yang, Liang Chen, Helan Hu, Runxin Xu, Qingxiu Dong, Ce Zheng, Shanghaoran Quan, Wen Xiao, Ge Zhang, Daoguang Zan, Keming Lu, Bowen Yu, Dayiheng Liu, Zeyu Cui, Jian Yang, Lei Sha, Houfeng Wang, Zhifang Sui, Peiyi Wang, Tianyu Liu, Baobao Chang
Preprint.
PDF Repo Page

LLM critics help catch bugs in mathematics: Towards a better mathematical verifier with natural language feedback

Bofei Gao, Zefan Cai, Runxin Xu, Peiyi Wang, Ce Zheng, Runji Lin, Keming Lu, Dayiheng Liu, Chang Zhou, Wen Xiao, Junjie Hu, Tianyu Liu, Baobao Chang
ACL 2025 (Findings) - Long Paper
PDF Code

PyramidKV: Dynamic KV Cache Compression based on Pyramidal Information Funneling

Zefan Cai, Yichi Zhang, Bofei Gao, Yuliang Liu, Yucheng Li, Tianyu Liu, Keming Lu, Wayne Xiong, Yue Dong, Junjie Hu, Wen Xiao
Preprint.
PDF Page Code Twitter Article 1 Article 2 Article 3 Article 4

Mitigating Language-Level Performance Disparity in mPLMs via Teacher Language Selection and Cross-lingual Self-Distillation

Haozhe Zhao*, Zefan Cai*, Shuzheng Si, Liang Chen, Yufeng He, Kaikai An, Baobao Chang
NAACL 2024 (Main) - Long Paper
PDF Code

Improving Event Definition Following For Zero-Shot Event Detection

Zefan Cai*, Po-Nien Kung*, Ashima Suvarna, Mingyu Derek Ma, Hritik Bansal, Baobao Chang, P Jeffrey Brantingham, Wei Wang, Nanyun Peng
ACL 2024 (Main) - Long Paper
PDF

PCA-Bench: Evaluating Multimodal Large Language Models in Perception-Cognition-Action Chain

Liang Chen, Yichi Zhang, Shuhuai Ren, Haozhe Zhao, Zefan Cai, Yuchi Wang, Peiyi Wang, Xiangdi Meng, Tianyu Liu, Baobao Chang
ACL 2024 (Main) - Long Paper
PDF Code Data

VeCAF: Vision-language Collaborative Active Finetuning with Training Objective Awareness

Rongyu Zhang*, Zefan Cai*, Huanrui Yang*, Zidong Liu, Denis A Gudovskiy, Tomoyuki Okuno, Yohei Nakata, Kurt Keutzer, Baobao Chang, Yuan Du, Li Du, Shanghang Zhang
ACM Multimedia 2024
PDF

MMICL: Empowering Vision-language Model with Multi-Modal In-Context Learning

Haozhe Zhao*, Zefan Cai*, Shuzheng Si*, Xiaojian Ma, Kaikai An, Liang Chen, Zixuan Liu, Sheng Wang, Wenjuan Han, Baobao Chang
ICLR 2024 (Poster)
PDF Code

Improving the Robustness of Distantly-Supervised Named Entity Recognition via Uncertainty-Aware Teacher Learning and Student-Student Collaborative Learning

Shuzheng Si, Helan Hu, Haozhe Zhao, Shuang Zeng, Kaikai An, Zefan Cai, Baobao Chang
ACL 2024 (Findings) - Long Paper
PDF

Towards end-to-end embodied decision making via multi-modal large language model: Explorations with gpt4-vision and beyond

Liang Chen, Yichi Zhang, Shuhuai Ren, Haozhe Zhao, Zefan Cai, Yuchi Wang, Peiyi Wang, Tianyu Liu, Baobao Chang
NeurIPS 2023 Foundation Models for Decision Making (FMDM) Workshop
PDF Code Data

Human-in-the-loop through chain-of-thought

Zefan Cai, Baobao Chang, Wenjuan Han
Preprint.
PDF

Large language models are not fair evaluators

Peiyi Wang, Lei Li, Liang Chen, Zefan Cai, Dawei Zhu, Binghuai Lin, Yunbo Cao, Qi Liu, Tianyu Liu, Zhifang Sui
ACL 2024 (Main) - Long Paper
PDF Code

DialogVCS: Robust Natural Language Understanding in Dialogue System Upgrade

Zefan Cai*, Xin Zheng*, Tianyu Liu*, Xu Wang, Haoran Meng, Jiaqi Han, Gang Yuan, Binghuai Lin, Baobao Chang, Yunbo Cao
NAACL 2024 (Main) - Long Paper
PDF

Diffcap: Exploring continuous diffusion on image captioning

Yufeng He*, Zefan Cai*, Xu Gan, Baobao Chang
Preprint.
PDF

SANTA: Separate Strategies for Inaccurate and Incomplete Annotation Noise in Distantly-Supervised Named Entity Recognition

Shuzheng Si*, Zefan Cai*, Shuang Zeng, Guoqiang Feng, Jiaxing Lin, Baobao Chang
ACL 2023 (Findings) - Long Paper
PDF

Compositional task representations for large language models

Nan Shao*, Zefan Cai*, Chonghua Liao, Yanan Zheng, Zhilin Yang
ICLR 2023 (Poster)
PDF Code