Capture Notes

Paper on trapping LLM model extraction attacks with knowledge honeypots.

AI security relevance:

• Relevant to model extraction detection/defense and deception-based protection.
• Can be connected to existing model extraction notes and Honeyquest.