HN
login
We benchmarked our prompt-injection detector against OWASP's LLM Topcerbereag.site

91.5% recall at sub-millisecond latency from regex alone, 98.1% with a ML layer added — and the false positives that come with it. Full methodology, mapped to the OWASP Top 10 for LLM Applications.

1 points•Dikesa_27•4 days ago•1 comment•
#llm#injection#prompt#detector#benchmarked#owasps#injection detector#prompt injection

1 comment

textcortex1 day ago
Can you also benchmark this one too? We recently trained it and it works well in production: https://jays.fyi/blog/open-prompt-injection-scanner-for-ai-a...

Read the full thread on Hacker News →

Related stories