profile photo

Zhaonan Li

I am a third-year Ph.D. student in Computer Science at ASU SCAI, advised by Prof. Ben Zhou.

My research focuses on post-training of large language and vision-language models, including reinforcement learning and distillation for reasoning, compute-efficient training and inference, and robust multimodal reasoning.

Previously, I was an applied research intern at Capital One and a research intern at the National Renewable Energy Laboratory (NREL). I received my B.S. (summa cum laude) in Computer Science and Applied Mathematics from Brandeis University and my M.S. in Computer Science from Columbia University.

I am looking for research internships for Summer 2027.

Email  /  Google Scholar  /  GitHub  /  LinkedIn  /  CV

News

Oct 2026

Our paper on adaptive decoding for token-efficient RL, from my Capital One internship, was accepted at the NeurIPS 2026 Workshop on Transitioning from Pre-training to Post-training.

Sep 2026

Consistency-Driven Learning for Proof-Carrying Reasoning was accepted at AACL 2026.

Aug 2026

Don’t Just Look, Intervene was accepted to Findings of EMNLP 2026.

Jun 2026

Started a summer internship at Capital One as an Applied Research Intern.

Apr 2026

VisAnalog was accepted as an oral at the CVPR 2026 Workshop on Visual Concepts.

Aug 2025

QA-LIGN was accepted to Findings of EMNLP 2025.

Jan 2025

One paper accepted at ICLR 2025 and one at NAACL 2025.

Publications

* equal contribution

Adaptive Decoding Methods for Token-Efficient Reinforcement Learning

Adaptive Decoding Methods for Token-Efficient Reinforcement Learning


Zhaonan Li*, Zelei Cheng*, Amritansh Mishra*, Chia-Hsuan Lee, Mann Patel, Sambit Sahu, Ben Zhou, William M. Campbell
NeurIPS 2026 Workshop on Transitioning from Pre-training to Post-training
arXiv

Work done during my internship at Capital One.

Consistency-Driven Learning for Proof-Carrying Reasoning

Consistency-Driven Learning for Proof-Carrying Reasoning


Xiao Ye, Zhaonan Li, Jacob Dineen, Shijie Lu, Shaswat Shrivastava, Avneet Ahuja, Ming Shen, Zhikun Xu, Ben Zhou
AACL 2026
arXiv

The arXiv version is titled "CC-LEARN: Cohort-based Consistency Learning".

Don't Just Look, Intervene: Perturbation Based Region Labeling for VQA Images

Don't Just Look, Intervene: Perturbation Based Region Labeling for VQA Images


Marko Jojic, Zhaonan Li, Ben Zhou
Findings of EMNLP 2026
arXiv
VisAnalog: A Diagnostic Suite for Visual Concept Transfer on Natural Images

VisAnalog: A Diagnostic Suite for Visual Concept Transfer on Natural Images


Zhaonan Li, Kyle R. Chickering, Bangzheng Li, Jacob Dineen, Xiao Ye, Zhikun Xu, Shijie Lu, Yuxi Huang, Ming Shen, Bach Nguyen, Jaya Adithya Pavuluri, Mau Son Nguyen, Sanika Chavan, Ngoc Minh Thu Le, Muhao Chen, Ben Zhou
CVPR 2026 Workshop on Visual Concepts Oral
arXiv / dataset
QA-LIGN: Aligning LLMs through Constitutionally Decomposed QA

QA-LIGN: Aligning LLMs through Constitutionally Decomposed QA


Jacob Dineen, Aswin RRV, Qin Liu, Zhikun Xu, Xiao Ye, Ming Shen, Zhaonan Li, Shijie Lu, Chitta Baral, Muhao Chen, Ben Zhou
Findings of EMNLP 2025
arXiv / paper
ToW: Thoughts of Words Improve Reasoning in Large Language Models

ToW: Thoughts of Words Improve Reasoning in Large Language Models


Zhikun Xu*, Ming Shen*, Jacob Dineen, Zhaonan Li, Xiao Ye, Shijie Lu, Aswin RRV, Chitta Baral, Ben Zhou
NAACL 2025
arXiv / paper / code
SysCaps: Language Interfaces for Simulation Surrogates of Complex Systems

SysCaps: Language Interfaces for Simulation Surrogates of Complex Systems


Patrick Emami, Zhaonan Li*, Saumya Sinha*, Truc Nguyen*
ICLR 2025
arXiv / code

Under Review & Preprints

Decoupled Reasoning and Perception for Generalizable Visual Reasoning

Decoupled Reasoning and Perception for Generalizable Visual Reasoning


Zhaonan Li*, Shijie Lu*, Fei Wang, Jacob Dineen, Xiao Ye, Zhikun Xu, Siyi Liu, Young Min Cho, Bangzheng Li, Daniel Chang, Kenny Nguyen, Qizheng Yang, Muhao Chen, Ben Zhou
Under review at CVPR 2027
arXiv

The arXiv version is titled "Unbiased Visual Reasoning with Controlled Visual Inputs".

Beyond Entropy: Token-Level Routing via Answer Impact

Beyond Entropy: Token-Level Routing via Answer Impact


Zhaonan Li*, Xiao Ye*, Yunyi Zhang, Ben Zhou
Under review at NAACL 2027
Self-Improving Vision-Language Models through Decoupled Self-Questioning

Self-Improving Vision-Language Models through Decoupled Self-Questioning


Zhaonan Li, Ben Zhou
Under review at NAACL 2027
How Well Do LLMs Reason with Noisy Evidence? An Active Visual Reasoning Benchmark

How Well Do LLMs Reason with Noisy Evidence? An Active Visual Reasoning Benchmark


Bach Nguyen, Zhaonan Li, Mau Son Nguyen, Sanika Chavan, Nilay Kumar, Hong Anh Nguyen, Khoa Vo, Ben Zhou
Under review at NAACL 2027
arXiv
Evaluating Medical LLMs by Levels of Autonomy: A Survey Moving from Benchmarks to Applications

Evaluating Medical LLMs by Levels of Autonomy: A Survey Moving from Benchmarks to Applications


Xiao Ye, Jacob Dineen, Zhaonan Li, Zhikun Xu, Weiyu Chen, Shijie Lu, Yuxi Huang, Ming Shen, Phu Tran, Ji-Eun Irene Yum, Muhammad Ali Khan, Muhammad Umar Afzal, Irbaz Bin Riaz, Ben Zhou
arXiv preprint
arXiv
Zero-Knowledge Zero-Shot Learning for Novel Visual Category Discovery

Zero-Knowledge Zero-Shot Learning for Novel Visual Category Discovery


Zhaonan Li, Hongfu Liu
arXiv preprint
arXiv

Design and source code from Jon Barron's website