On: Garak, LLM Vulnerability Scanner

[ivanbalepin]

I'd imagine there is a big difference between ML-based heuristic detection for traditional AV and testing for malicious prompts, no? Like, why can't BofA kill Paypal difference.