Product
Help
Pricing
About
Jobs
Log in
Get started
Author
Add to collection
Zenghao Duan
Observed name:
Duan, Zenghao
Institutions:
Chinese Academy of Sciences,
Dalian University of Technology,
Institute of Computing Technology,
University of Chinese Academy of Sciences
Observed institutions:
Chinese Academy of Sciences,
Dalian University of Technology,
Institute of Computing Technology,
University of Chinese Academy of Sciences
OpenAlex ID:
A5122941204
The Evolution of Thought: Tracking LLM Overthinking via Reasoning Dynamics Analysis
2026
·
Zihao Wei
, Liang Pang
, et al.
·
Proceedings of the conference - Association for Computational Linguistics. Meeting
·
1
PDF
SkillAtlas: An Attack Trace Library for Agent Skills
2026
·
Tian, Yuxin
, Duan, Zenghao
, et al.
·
arXiv (Cornell University)
PDF
Projecting Out the Malice: A Global Subspace Approach to LLM Detoxification
2026
·
Association for Computational Linguistics 2026
, Cheng, Xueqi
, et al.
·
Underline Science Inc.
The Evolution of Thought: Tracking LLM Overthinking via Reasoning Dynamics Analysis
2026
·
Association for Computational Linguistics 2026
, Cheng, Xueqi
, et al.
·
Underline Science Inc.
Circular Reasoning: Understanding Self-Reinforcing Loops in Large Reasoning Models
2026
·
Duan, Zenghao
, Pang, Liang
, et al.
·
arXiv (Cornell University)
PDF
Projecting Out the Malice: A Global Subspace Approach to LLM Detoxification
2026
·
Duan, Zenghao
, Yin, Zhiyi
, et al.
·
arXiv (Cornell University)
PDF
SkillAttack: Automated Red Teaming of Agent Skills through Attack Path Refinement
2026
·
Zenghao Duan
, Yuxin Tian
, et al.
·
Conference on Empirical Methods in Natural Language Processing (EMNLP)
PDF
Latent-GRPO: Group Relative Policy Optimization for Latent Reasoning
2026
·
Jingcheng Deng
, Zihao Wei
, et al.
·
Conference on Empirical Methods in Natural Language Processing (EMNLP)
PDF
Dynamic Rollout Editing for Reducing Overthinking in RL-Trained Reasoning Models
2026
·
Zihao Wei
, Wenjie Shi
, et al.
·
Conference on Empirical Methods in Natural Language Processing (EMNLP)
PDF
Projecting Out the Malice: A Global Subspace Approach to LLM Detoxification
2026
·
Zenghao Duan
, Zhiyi Yin
, et al.
·
Proceedings of the conference - Association for Computational Linguistics. Meeting
PDF
Latent-SFT: LLM Latent Reasoning as Chain of Superposition
2026
·
Jingcheng Deng
, Liang Pang
, et al.
·
Conference on Empirical Methods in Natural Language Processing (EMNLP)
PDF
Show more works
View as search
Works count:
17
Citation count:
2
H-index:
1
I10-index:
Year
Topic
Type
