Shypd.aiThe largest open index of AI toolsPurpleLlama is an open-source project that provides tools and evaluations to assess and improve LLM security. It helps developers build responsibly with open generative AI models by offering cybersecurity and input/output safeguards.
Trending
About
PurpleLlama is an umbrella project from Meta that consolidates tools and evaluations designed to enhance the security and responsible development of open generative AI models. It adopts a 'purple teaming' approach, combining offensive (red team) and defensive (blue team) strategies to mitigate potential risks. Key components include Llama Guard, a series of high-performance input and output moderation models for detecting violating content and preventing malicious code execution; Prompt Guard, which protects against prompt injection and jailbreaking attacks; and Code Shield, for filtering insecure code produced by LLMs. The project also features CyberSec Eval benchmarks to quantify LLM cybersecurity risks and evaluate the frequency of insecure code suggestions, supporting the development of safer AI systems.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending
