SubmissionNumber#=%=#29
FinalPaperTitle#=%=#CAPTURE: Context-Aware Prompt Injection Testing and Robustness Enhancement
ShortPaperTitle#=%=#
NumberOfPages#=%=#13
CopyrightSigned#=%=#GAURI PRADIPKUMAR KHOLKAR
JobTitle#==#
Organization#==#PURE STORAGE
3rd Floor, Pure Storage India Pvt Ltd, 1 Sobha, Shanthala Nagar, Ashok Nagar, Bengaluru, Karnataka 560001
Abstract#==#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.
Author{1}{Firstname}#=%=#Gauri
Author{1}{Lastname}#=%=#Kholkar
Author{1}{Username}#=%=#kgauri
Author{1}{Email}#=%=#gkholkar@purestorage.com
Author{1}{Affiliation}#=%=#Pure Storage
Author{2}{Firstname}#=%=#Ratinder
Author{2}{Lastname}#=%=#Ahuja
Author{2}{Email}#=%=#rahuja@purestorage.com
Author{2}{Affiliation}#=%=#Ratinder Ahuja

==========
èéáğö