WhileBug

Department of Computer Science

University of California, Los Angeles

I’m Peiran Wang (王沛然 in Chinese, you can speak as “Pay-than, Wang”), a CS PhD student in UCLA. I am currently interested in general AI Security (including LLM security, privacy, etc. Especially system security for LLM recently), Machine Learning System (MLSys), usable security, fraud detection and program analysis, etc. I graduated from Sichuan University in Cybersecurity Talented Class in 2022 (18/172). During my undergraduate, I got National Scholarship (3/172), and got GPA of 3.92 (rank 2/172). I’m currently a CS PhD student in UCLA, under the supervision of Prof. Yuan Tian.

research areas

AI Agent Security

Securing autonomous AI agents against adversarial manipulation, prompt injection, and unintended behaviors in real-world deployments.

Interpretable AI Security

Leveraging interpretability and explainability techniques to understand, diagnose, and mitigate vulnerabilities in AI systems.

Usable Security of AI

Designing intuitive security mechanisms and interfaces that help users safely interact with and configure AI systems.

AI Misuse Measurement

Developing methods to quantify and detect malicious applications of AI, from deepfakes to automated cyber attacks.

AI Society Security

Investigating the societal impacts of AI security risks, including misinformation, surveillance, and governance challenges.

AI for Security

Applying AI techniques to strengthen cybersecurity defenses, including threat detection, vulnerability analysis, and automated response.

news

Feb 01, 2026	Our lab is selected as one of the 10 teams for Amazon Nova AI Challenge! I serve as one of the team leaders!
Dec 22, 2025	Receive PhD research internship offer from NVIDIA, see you in Santa Clara!
Mar 30, 2025	Got PhD offer from CS@UCLA!
Jan 22, 2025	Finished the research internship in ByteDance.
Oct 19, 2024	Moderator is awarded CCS Distinguished Paper Award at CCS 2024.

latest posts

Mar 22, 2026	猫猫咪咪出国记
May 14, 2024	Google Gemini updates: Flash 1.5, Gemma 2 and Project Astra
Apr 23, 2022	Displaying External Posts on Your al-folio Blog

selected publications

The Landscape of Prompt Injection Threats in LLM Agents: From Taxonomy to Analysis

Peiran Wang, Xinfeng Li, Chong Xiang, and 5 more authors

arXiv preprint, 2026

Bib HTML

@article{wang2026pisok,
  title = {The Landscape of Prompt Injection Threats in LLM Agents: From Taxonomy to Analysis},
  author = {Wang, Peiran and Li, Xinfeng and Xiang, Chong and Zhang, Jinghuai and Li, Ying and Zhang, Lixia and Wang, Xiaofeng and Tian, Yuan},
  year = {2026},
  journal = {arXiv preprint},
}

AgentArmor: Enforcing Program Analysis on Agent Runtime Trace to Defend Against Prompt Injection

Peiran Wang, Yang Liu, Yunfei Lu, and 6 more authors

arXiv preprint, 2025

Bib HTML

@article{wang2025agentarmor,
  title = {AgentArmor: Enforcing Program Analysis on Agent Runtime Trace to Defend Against Prompt Injection},
  author = {Wang, Peiran and Liu, Yang and Lu, Yunfei and Cai, Yifeng and Chen, Hongbo and Yang, Qingyou and Zhang, Jie and Hong, Jue and Wu, Ye},
  year = {2025},
  journal = {arXiv preprint},
}

Large Language Model-based Data Science Agent: A Survey

Ke Chen, Peiran Wang, Yaoning Yu, and 2 more authors

Transactions on Machine Learning Research, 2025

Bib HTML

@article{chen2025dsagentsurvey,
  title = {Large Language Model-based Data Science Agent: A Survey},
  author = {Chen, Ke and Wang, Peiran and Yu, Yaoning and Zhan, Xianyang and Wang, Haohan},
  year = {2025},
  journal = {Transactions on Machine Learning Research},
}

What Are Models Thinking About? Understanding Large Language Model Hallucinations "Psychology" Through Model Inner State Analysis

Peiran Wang, Yang Liu, Yunfei Lu, and 2 more authors

arXiv preprint, 2025

Bib HTML

@article{wang2025haluprobe,
  title = {What Are Models Thinking About? Understanding Large Language Model Hallucinations "Psychology" Through Model Inner State Analysis},
  author = {Wang, Peiran and Liu, Yang and Lu, Yunfei and Hong, Jue and Wu, Ye},
  year = {2025},
  journal = {arXiv preprint},
}

Moderator: Moderating Text-to-Image Diffusion Models through Fine-grained Context-based Policies

Peiran Wang, Qiyu Li, Longxuan Yu, and 3 more authors

In 31st ACM Conference on Computer and Communications Security (CCS), 2024

Awarded Bib HTML

CCS Distinguished Paper Award

@inproceedings{wang2024moderator,
  title = {Moderator: Moderating Text-to-Image Diffusion Models through Fine-grained Context-based Policies},
  author = {Wang, Peiran and Li, Qiyu and Yu, Longxuan and Wang, Ziyao and Li, Ang and Jin, Haojian},
  year = {2024},
  booktitle = {31st ACM Conference on Computer and Communications Security (CCS)},
}