Boxin Wang

Senior Research Scientist
NVIDIA

Contact: boxinw@nvidia.com
[Google Scholar] [GitHub] [Linkedln]

I am a Senior Research Scientist at NVIDIA. I obtained my Ph.D. degree from the Computer Science department of University of Illinois, Urbana-Champaign (UIUC). During my Ph.D., I was supervised by Prof. Bo Li.

I enjoy working on practical research: simple ideas that hold up at scale and make their way into frontier model development. My research interests mainly lie in LLM post-training, including:

Long-horizon agentic coding
Reasoning, alignment, and trustworthiness
Multi-modal language modeling

News

[2026/09] Our work Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models is accepted to NeurIPS 2026 as an Oral presentation (oral acceptance rate: 0.36%, 112 out of 30,709 submissions) 🎉 See you in Sydney, Australia!
[2026/03] Our work Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation is out! Honored to be part of the core contributos to build frontier LLM with Gold Medal 🏅 level performance on IMO, IOI, ICPC World Finals.
[2024/09] Our work DecodingTrust is awarded the winner of Cybersecurity Award in the track of "Best Machine Learning and Security Paper" 🎉
[2023/12] Our work DecodingTrust is awarded NeurIPS 2023 outstanding paper. 🎉

More news

[2025/04] Our work Nemotron-H: A Family of Accurate and Efficient Hybrid Mamba-Transformer Models is out! Honored to be part of the core contributors to build VLM for Nemotron-H
[2024/03] Our work Cosmos-Reason1: From Physical Common Sense To Embodied Reasoning is out! Honored to be part of the core contributors for base VLM training of Cosmos-Reason1
[2024/09] Our work NVLM: Open Frontier-Class Multimodal LLMs is out! Feel free to reach out to me if you are interested in our work 🎉
[2023/10] I passed my Ph.D. defense! 🎉
[2023/05] I started my internship at NVIDIA. Look forward to seeing you in the Bay Area!
[2023/05] I am organizing Trustworthy and Reliable Large-Scale Machine Learning Models workshop at ICLR 2023.

Publications

2026

Unified Audio Intelligence Without Regressing on Text Intelligence 2026

Zhifeng Kong*, Sang-gil Lee*, Jaehyeon Kim*, Boxin Wang*, Zihan Liu, Sungwon Kim, Yang Chen, Arushi Goel, Rajarshi Roy, Wenliang Dai, Zhuolin Yang, Yangyi Chen, Dongfu Jiang, Sreyan Ghosh, Tuomas Rintamaki, Andrew Tao, Jonathan Raiman, Mohammad Shoeybi, Bryan Catanzaro, Wei Ping* [PDF][Checkpoints & Demo]

Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation 2026

Zhuolin Yang*, Zihan Liu*, Yang Chen*, Wenliang Dai*, Boxin Wang*, Sheng-Chieh Lin, Chankyu Lee, Yangyi Chen, Dongfu Jiang, Jiafan He, Renjie Pi, Grace Lam, Nayeon Lee, Alexander Bukharin, Mohammad Shoeybi, Bryan Catanzaro, Wei Ping* [PDF][Models & Datasets][Blog]

From 128K to 4M: Efficient Training of Ultra-Long Context Large Language Models ACL 2026 (Findings)

Chejian Xu, Wei Ping, Peng Xu, Zihan Liu, Boxin Wang, Mohammad Shoeybi, Bo Li, Bryan Catanzaro [PDF][Models]

2025

Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models NeurIPS 2026 (Oral)

Boxin Wang*, Chankyu Lee*, Nayeon Lee*, Sheng-Chieh Lin*, Wenliang Dai*, Yang Chen*, Yangyi Chen*, Zhuolin Yang*, Zihan Liu*, Mohammad Shoeybi, Bryan Catanzaro, Wei Ping [PDF][Models & Datasets][Blog]

Nemotron-H: A Family of Accurate and Efficient Hybrid Mamba-Transformer Models 2025

NVIDIA: Boxin Wang (core contributor), et al. [PDF][Blog]

Cosmos-Reason1: From Physical Common Sense To Embodied Reasoning 2025

NVIDIA: Boxin Wang (core contributor), et al. [PDF][Blog]

2024

NVLM: Open Frontier-Class Multimodal LLMs 2024

Wenliang Dai*, Nayeon Lee*, Boxin Wang*, Zhuoling Yang*, Zihan Liu, Jon Barker, Tuomas Rintamaki, Mohammad Shoeybi, Bryan Catanzaro, Wei Ping*. (equal contribution, listed alphabetically) [PDF][project]

RankRAG: Unifying Retrieval-Augmented Generation and Context Ranking in LLMs NeurIPS 2024

Yue Yu, Wei Ping, Zihan Liu, Boxin Wang, Jiaxuan You, Chao Zhang, Mohammad Shoeybi, Bryan Catanzaro [PDF]

InstructRetro: Instruction Tuning post Retrieval-Augmented Pretraining ICML 2024

Boxin Wang, Wei Ping, Lawrence McAfee, Peng Xu, Bo Li, Mohammad Shoeybi, Bryan Catanzaro. [PDF][code]

Can Public Large Language Models Help Private Cross-device Federated Learning? NAACL 2024 (Findings)

Boxin Wang, Yibo Jacky Zhang, Yuan Cao, Bo Li, H. Brendan McMahan, Sewoong Oh, Zheng Xu, Manzil Zaheer [PDF]

UNICORN: A Unified Causal Video-Oriented Language-Modeling Framework for Temporal Video-Language Tasks EMNLP 2024

Yuanhao Xiong, Yixin Nie, Haotian Liu, Boxin Wang, Jun Chen, Rong Jin, Cho-Jui Hsieh, Lorenzo Torresani, Jie Lei [PDF]

2023

DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models NeurIPS 2023 (Outstanding Paper)

Boxin Wang, Weixin Chen, Hengzhi Pei, Chulin Xie, Mintong Kang, Chenhui Zhang, Chejian Xu, Zidi Xiong, Ritik Dutta, Rylan Schaeffer, Sang T. Truong, Simran Arora, Mantas Mazeika, Dan Hendrycks, Zinan Lin, Yu Cheng, Sanmi Koyejo, Dawn Song, Bo Li. [PDF][website][code]

Shall We Pretrain Autoregressive Language Models with Retrieval? A Comprehensive Study EMNLP 2023

Boxin Wang, Wei Ping, Peng Xu, Lawrence McAfee, Zihan Liu, Mohammad Shoeybi, Yi Dong, Oleksii Kuchaiev, Bo Li, Chaowei Xiao, Anima Anandkumar, Bryan Catanzaro [PDF][code]

2022

Exploring the Limits of Domain-Adaptive Training for Detoxifying Large-Scale Language Models NeurIPS 2022

Boxin Wang, Wei Ping, Chaowei Xiao, Peng Xu, Mostofa Patwary, Mohammad Shoeybi, Bo Li, Anima Anandkumar, Bryan Catanzaro [PDF][code]

Improving Certified Robustness via Statistical Learning with Logical Reasoning NeurIPS 2022

Zhuolin Yang, Zhikuan Zhao, Boxin Wang, Jiawei Zhang, Linyi Li, Hengzhi Pei, Bojan Karlaš, Ji Liu, Heng Guo, Ce Zhang, Bo Li [PDF]

SemAttack: Natural Textual Attacks via Different Semantic Spaces NAACL 2022 (Findings)

Boxin Wang*, Chejian Xu*, Xiangyu Liu, Yu Cheng, Bo Li [PDF] [Code]

Certifying Out-of-Domain Generalization for Blackbox Functions ICML 2022

Maurice Weber, Linyi Li, Boxin Wang, Zhikuan Zhao, Bo Li, Ce Zhang [PDF]

2021

Adversarial GLUE: A Multi-Task Benchmark for Robustness Evaluation of Language Models NeurIPS 2021 (Oral)

Boxin Wang*, Chejian Xu*, Shuohang Wang, Zhe Gan, Yu Cheng, Jianfeng Gao, Ahmed Hassan Awadallah, Bo Li [PDF] [Dataset]

G-PATE: Scalable Differentially Private Data Generator via Private Aggregation
of Teacher Discriminators NeurIPS 2021

Yunhui Long*, Boxin Wang*, Zhuolin Yang, Bhavya Kailkhura, Aston Zhang, Carl A. Gunter, Bo Li [PDF] [Code]

DataLens: Scalable Privacy Preserving Training via Gradient Compression and Aggregation CCS 2021

Boxin Wang*, Fan Wu*, Yunhui Long*, Luka Rimanic, Ce Zhang, Bo Li [PDF] [Code]

Uncovering the Connections Between Adversarial Transferability and Knowledge Transferability ICML 2021

Kaizhao Liang*, Jacky Zhang*, Boxin Wang, Zhuolin Yang, Sanmi Koyejo, Bo Li [PDF] [Code]

Counterfactual Adversarial Learning with Representation Interpolation EMNLP 2021

Wei Wang, Boxin Wang, Ning Shi, Jinfeng Li, Bingyu Zhu, Xiangyu Liu, Rong Zhang [PDF] [Code]

Incorporating External POS Tagger for Punctuation Restoration Interspeech 2021

Ning Shi, Wei Wang, Boxin Wang, Jinfeng Li, Xiangyu Liu, Zhouhan Lin [PDF][Code]

InfoBERT: Improving Robustness of Language Models from An Information Theoretic Perspective ICLR 2021

Boxin Wang, Shuohang Wang, Yu Cheng, Zhe Gan, Ruoxi Jia, Bo Li, Jingjing Liu [PDF][Code]

2020

T3: Tree-Autoencoder Regularized Adversarial Text Generation for Targeted Attack EMNLP 2020

Boxin Wang, Hengzhi Pei, Boyuan Pan, Qian Chen, Shuohang Wang and Bo Li [PDF][Code]

Reinforcement-Learning based Portfolio Management with Augmented Asset Movement
Prediction States AAAI 2020

Yunan Ye, Hengzhi Pei, Boxin Wang, Pin-Yu Chen, Yada Zhu, Jun Xiao, Bo Li [PDF]

2019

Efficient task-specific data valuation for nearest neighbor algorithms PVLDB 2019

Ruoxi Jia*, David Dao*, Boxin Wang, Frances Ann Hubis, Nick Hynes, Nezihe Merve Gürel, Bo Li, Ce Zhang, Costas J. Spanos, Dawn Song [PDF]

Towards efficient data valuation based on the shapley value AISTATS 2019

Ruoxi Jia*, David Dao*, Boxin Wang, Frances Ann Hubis, Nick Hynes, Nezihe Merve Gürel, Bo Li, Ce Zhang, Dawn Song, Costas J. Spanos [PDF]

Personal

I like photographing. I put my photos on 500px.