Hands-On AI Security: Red Teaming Piper with 5 Prompt Injection Vectors
How to probe and break local LLMs using Direct Injection, Roleplay, Obfuscation, and Authority Claims, demonstrating why system instructions fail as security boundaries.
How to probe and break local LLMs using Direct Injection, Roleplay, Obfuscation, and Authority Claims, demonstrating why system instructions fail as security boundaries.