Prompt injection

From Rice Wiki
Revision as of 20:28, 23 May 2024 by Rice (talk | contribs)


A prompt injection attack involves a user injecting a malicious instruction in an LLM-integrated application, in which user input was intended to act as only data.

Vulnerability

Prompt injection exploits the single-channel nature of LLM's, where user prompts and system prompts are simply concatenated together and processed.

Defense strategies