Lesson60 seconds
What is Jailbreak?
Tricks that try to make an AI ignore its safety rules.
Step 1 of 9
Simple Definition
A jailbreak is a prompting technique aimed at bypassing a model's safety policies to elicit disallowed outputs.
Explanation level
Tap a persona to hear the same idea in a totally different voice.
Try “Like I'm 5™” for a picture-book version - not the grown-up wording above.
Notebook
Your notes
Notes save automatically on this device
Knowledge graph
Done Here Next