Title: research | WhileBug
Description: Our research spans six key areas in AI security.
Keywords:
Domain: whilebug.github.io
| None | default-src 'self'; script-src 'self' 'unsafe-inline' https:; style-src 'self' 'unsafe-inline' https:; img-src 'self' data: https:; font-src 'self' data: https:; media-src 'self' https:; frame-src 'self' https:; connect-src 'self' https:; |
| author | Peiran Wang |
Links:
| WhileBug | https://whilebug.github.io/ |
| about | https://whilebug.github.io/ |
| blog | https://whilebug.github.io/blog/ |
| publications | https://whilebug.github.io/publications/ |
| research (current) | https://whilebug.github.io/research/ |
| CV | https://whilebug.github.io/cv/ |
| [Arxiv-2026] PI-SoK: The Landscape of Prompt Injection Threats in LLM Agents | https://arxiv.org/abs/2602.10453 |
| [Arxiv-2025] AgentArmor: Enforcing Program Analysis on Agent Runtime Trace to Defend Against Prompt Injection | https://www.arxiv.org/abs/2508.01249 |
| [Arxiv-2024] PIB-Bench: Prompt Injection Benchmark for Foundation Model Integrated Systems | https://openreview.net/forum?id=MsRdq0ePTR |
| [Arxiv-2025] HaluProbe: What Are Models Thinking About? Understanding LLM Hallucinations Through Model Inner State Analysis | https://arxiv.org/abs/2502.13490 |
| [Arxiv-2025] Halu2Jail: From Hallucinations to Jailbreaks: Rethinking the Vulnerability of Large Foundation Models | https://www.arxiv.org/abs/2505.24232 |
| [CCS-2024] Moderator: Moderating Text-to-Image Diffusion Models through Fine-grained Context-based Policies | https://arxiv.org/abs/2408.07728 |
| [NAACL-2025] RePD: Defending Jailbreak Attacks Through a Retrieval-based Prompt Decomposition Process | https://aclanthology.org/2025.findings-naacl.16/ |
| [Arxiv-2026] MCPSmell: From Docs to Descriptions: Smell-Aware Evaluation of MCP Server Descriptions | https://arxiv.org/abs/2602.18914 |
| [ICLR-2026] TrustGen: A Platform of Dynamic Benchmarking on the Trustworthiness of Generative Foundation Models | https://openreview.net/forum?id=Fcf5fLmaeG |
| [NAACL-2025] CVE-Bench: Benchmarking LLM-based Software Engineering Agent’s Ability to Repair Real-World CVE Vulnerabilities | https://aclanthology.org/2025.naacl-long.212/ |
| Jekyll | https://jekyllrb.com/ |
| al-folio | https://github.com/alshedivat/al-folio |
| GitHub Pages | https://pages.github.com/ |
Viewport: width=device-width, initial-scale=1, shrink-to-fit=no