ai security benchmarking

CrowdStrike and Meta have launched CyberSOCEval, an open-source benchmarking suite to evaluate large language models' effectiveness in critical security tasks.…