Jun Bai
Senior Research Engineer · BIGAI NLCo Lab
About me
I received my Ph.D. degree from the School of Computer Science and Engineering at Beihang University, China. I completed my doctoral studies under the supervision of Prof. Wenge Rong and Prof. Chuantao Yin.
My research primarily focuses on building Trustworthy AI systems. My current research themes include:
- Belief memory in agents: learning, retrieving, and revising beliefs over time.
- Improving LLMs’ faithful use of context for evidence-grounded generation and reasoning.
- Interpretability of LLMs and agents: explaining internal reasoning, memory use, and decision-making.
I am always looking for self-motivated interns. Please feel free to drop me an email if interested.
News
Belief Context Graph (BCG) has been accepted by the NeurIPS 2026 Workshop on Resource-Aware Agentic AI. Code is open source; the paper is forthcoming.
OneMillion-Bench has been accepted by NeurIPS 2026 (Evaluations and Datasets Track).
CircuitLens has been accepted by EMNLP 2026 Findings.
Native Parallel Reasoner has been accepted by ICML 2026.
RouterLens, TongSearch-QR, and CogAtom are accepted by EMNLP 2025.
Selected publications
All publications ↗Belief Context Graph (BCG)
Open-source belief-aware context infrastructure for long-horizon agents. Paper forthcoming.
NeurIPS Workshop on Resource-Aware Agentic AI 2026
OneMillion-Bench: How Far are Language Agents from Human Experts?
Qianyu Yang*, Yang Liu*, Jiaqi Li*, Jun Bai*, Hao Chen, Kaiyuan Chen, Tiliang Duan, Jiayun Dong, Xiaobo Hu, Zixia Jia, Yang Liu, Tao Peng, Yixin Ren, Ran Tian, Zaiyuan Wang, Yanglihong Xiao, Gang Yao, Lingyue Yin, Ge Zhang, Chun Zhang, Jianpeng Jiao, Zilong Zheng, Yuan Gong
NeurIPS 2026Evaluations and Datasets Track
* equal contribution
CircuitLens: Reasoning Circuits as Data Selection Signals for Reinforcement Learning with Verifiable Rewards
Zhuofan Chen, Ziqian Jiao, Yikai Cui, Zhixin Cai, Jun Bai, Wenge Rong
EMNLP Findings 2026
Understanding and Leveraging the Expert Specialization of Context Faithfulness in Mixture-of-Experts LLMs
Jun Bai, Minghao Tong, Yang Liu, Zixia Jia, Zilong Zheng
EMNLP 2025Top 2%
Reinforced Query Reasoners for Reasoning-intensive Retrieval Tasks
Xubo Qin, Jun Bai, Jiaqi Li, Zixia Jia, Zilong Zheng
EMNLP 2025
Leveraging Estimated Transferability over Human Intuition for Model Selection in Text Ranking
Jun Bai, Zhuofan Chen, Zhenzi Li, Hanhua Hong, Jianfei Zhang, Chen Li, Chenghua Lin, Wenge Rong
EMNLP 2024
How to Determine the Most Powerful Pre-trained Language Model without Brute Force Fine-tuning? An Empirical Survey
Jun Bai, Xiaofeng Zhang, Chen Li, Hanhua Hong, Xi Xu, Chenghua Lin, Wenge Rong
EMNLP Findings 2023
