Design effective LLM prompts to avoid prompt injections

RocketChat / Apps.RC.AI.Programmer

Generate short code snippets in various languages using LLMs within Rocket.Chat. Share directly or create pull requests on GitHub.

1 stars 0 forks source link

Design effective LLM prompts to avoid prompt injections #21

Closed RyanbowZ closed 1 month ago

RyanbowZ commented 2 months ago

For the prompt engineering part, the previous prompts are very ineffective and prone to attack. We should set a more effective LLM prompts to avoid improper injections and ensure the app function as expected.