Classify and filter LLM content into six safety categories.
npx skills add https://github.com/jonnabio/ace-framework --skill llamaguard-jonnabio
Please help me install this Agent Skill. Skill: llamaguard Source: https://github.com/jonnabio/ace-framework/tree/main/.ace/packs/ai-research/llamaguard Command: npx skills add https://github.com/jonnabio/ace-framework --skill llamaguard-jonnabio
This Skill addresses the challenge of ensuring AI-generated content is safe and adheres to ethical standards, by integrating Meta's LlamaGuard moderation model.
Install the required libraries and use the 'moderate' function to check user input or LLM responses.
High-intent search queries and answers about installing and using this skill.
You can moderate LLM input and output for violence, hate, and sexual content by integrating Meta's LlamaGuard model. It classifies and filters text across six predefined safety categories to ensure AI content compliance.
Meta LlamaGuard is a specialized moderation model used for AI content safety. It classifies and filters prompts and responses to prevent unsafe content like criminal planning, weapons, and self-harm from being generated.
Yes, LlamaGuard works with vLLM and HuggingFace transformers. It requires the transformers, torch, and vllm libraries to run the moderation model and classify content into safety categories.
Yes, you can use this moderation model with Sagemaker and NeMo Guardrails. The model is designed for integration across various AI platforms to filter unsafe LLM responses.
To classify AI generated text into predefined safety categories, use the LlamaGuard moderate function. It evaluates user input and LLM responses against six safety categories including violence, hate, and substances.
Limitations include its dependency on specific libraries like transformers, torch, and vllm. The model only classifies content into six predefined safety categories, which may limit custom moderation requirements.
Orchestra-ResearchClassifies LLM inputs and outputs into six safety categories using Meta's LlamaGuard moderation model.
WaterplanAIManages multi-account Google Workspace services via a CLI with orchestration and permissions.
OpenLAIRClassify and filter unsafe LLM inputs and outputs across six safety categories.
ricableClassify LLM inputs and outputs into six safety categories.
