Formalizing and Benchmarking Prompt Injection Attacks and Defenses

A prompt injection attack aims to inject malicious instruction/data into the input of an LLM-Integrated Application such that it produces results as an attacker desires. Existing works are limited to case studies. As a result, the literature lacks a systematic understanding of prompt injection attacks and their defenses. We aim to bridge the gap in this work. In particular, we propose a framewo…

Paper

Similar papers

© 2026 NYSGPT2525 LLC