2025/05/18 by Gauri Kholkar, Kholkar, Gauri, Ahuja, Ratinder · 1 citation
Computer Science · #Adversarial Robustness in Machine Learning #Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #FOS: Computer and information sciences #Security and Verification in Computing #Topic Modeling
paper · pdf · doi:10.48550/arxiv.2505.12368
openalex publication_date 2025/05/18 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Prompt injection remains a major security risk for large language models. However, the efficacy of existing guardrail models in context-aware settings remains underexplored, as they often rely on static attack benchmarks. Additionally, they have over-defense tendencies. We introduce CAPTURE, a novel context-aware benchmark assessing both attack detection and over-defense tendencies with minimal in-domain examples. Our experiments reveal that current prompt injection guardrail models suffer from high false negatives in adversarial cases and excessive false positives in benign scenarios, highlighting critical limitations. To demonstrate our framework's utility, we train CaptureGuard on our generated data. This new model drastically reduces both false negative and false positive rates on our context-aware datasets while also generalizing effectively to external benchmarks, establishing a path toward more robust and practical prompt injection defenses.