Benchmarking LLMs for Cybersecurity Applications
After announcing our joint work between Meta and CrowdStrike at LlamaCon in April, we finally released our new LLM benchmark for cybersecurity applications, CyberSOCEval, this month at Fal.Con. You can read our paper here.
The CyberSOCEval benchmark code is available as part of CyberSecEval 4, and the benchmark includes cybersecurity-specific data for analysis. In addition, there’s more information in our press release and in coverage by ZDNET.
Subscribe via RSS
