.svg)
What is CyberSecEval Harmful Content Attack?
This evaluation method tests AI systems using samples from the CyberSecEval dataset. It's specifically designed to assess how well AI agents can resist generating harmful or misleading content related to cybersecurity.
