5 papers
Prompt Overflow: What the Guardrail Inspects Is Not What the Model Infers
Yuanbo Zhou, Changjia Zhu, Junyu Wang +5
Guardrail models (a.k.a. safety checkers) are widely deployed to screen user inputs before they reach large language models (LLMs), serving as a primary defense against prompt inje…
PathFix: Automated Program Repair with Expected Path
Xu He, Shu Wang, Kun Sun
Automated program repair (APR) techniques are effective in fixing inevitable defects in software, enhancing development efficiency and software robustness. However, due to the diff…
ChainMarks: Securing DNN Watermark with Cryptographic Chain
Brian Choi, Shu Wang, Isabelle Choi +1
With the widespread deployment of deep neural network (DNN) models, dynamic watermarking techniques are being used to protect the intellectual property of model owners. However, re…
SentinelAgent: Graph-based Anomaly Detection in Multi-Agent Systems
Xu He, Di Wu, Yan Zhai +1
The rise of large language model (LLM)-based multi-agent systems (MAS) introduces new security and reliability challenges. While these systems show great promise in decomposing and…
TYPEPULSE: Detecting Type Confusion Bugs in Rust Programs
Hung-Mao Chen, Xu He, Shu Wang +2
Rust supports type conversions and safe Rust guarantees the security of these conversions through robust static type checking and strict ownership guidelines. However, there are in…