Llama Guard 3
Named by 3 kept items in this workspace.
2 items name it under tools and 1 under classification.technologies.
Search every workspace for Llama Guard 3. It matches the name as written, so it can find other things that share it.
The items that name it 3
Competing content-safety guardrail model (8B); used as baseline and as a target for plug-in augmentation with NSFA-trained classification heads (+17.6 F1 points)
series of input and output moderation models to detect common types of violating content, optimized to detect MLCommons standard hazards taxonomy, cyberattack responses, and malicious code output for LLM code interpreter environments
All technologies in this workspace.