Files
raccoon/injections.md
2023-03-02 20:58:40 +01:00

13 lines
527 B
Markdown

# Injection Methods
This is a list of ways, in which an LLM might get injected with a malicious prompt or payload.
+ Social Media
+ Twitter
+ [Smart Feed](https://github.com/SmartLever/SmartFeeds),
+ Email [llm-security](https://github.com/greshake/llm-security)
+ Websites [llm-security](https://github.com/greshake/llm-security)
+ Bing
+ [splendidbing](https://github.com/velocitatem/llm-cross-prompt-scripting/blob/main/splendidbing.md)
+ [Indirect Prompt Injection on Bing Chat](https://greshake.github.io/)