Announcement
Cybersecurity monitors for Kimi K3 and GLM 5.3 claimed to be 50x faster and 50x cheaper than an optimized LLM judge
Goodfire says external red-teaming from @farairesearch found its monitor cut universal jailbreak success to 0%.
TLDR
Goodfire says it built cybersecurity monitors for Kimi K3 and GLM 5.3 that are more accurate, 50x faster and 50x cheaper than an optimized LLM judge. It also says external red-teaming from @farairesearch found its monitor cut universal jailbreak success to 0%.
Combined views
—
11 Sources, first seen ago
