Meta AI Introduces CyberSecEval 2: A Novel Machine Learning Benchmark to Quantify LLM Security Risks and Capabilities
Marktechpost
MAY 1, 2024
Prior works to evaluate LLMs’ security properties include open benchmark frameworks and position papers proposing evaluation criteria. Meta researchers present CYBERSECEVAL 2, a benchmark for assessing LLMs security risks and capabilities, including prompt injection and code interpreter abuse testing.
Let's personalize your content