2 papers
cs.CR2026
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…
cs.CR2024
JNI Global References Are Still Vulnerable: Attacks and Defenses
Yi He, Yuan Zhou, Yacong Gu +4
System services and resources in Android are accessed through IPC based mechanisms. Previous research has demonstrated that they are vulnerable to the denial-of-service attack (DoS…