Llama-Guard-4-12B is Meta's latest open-weight safeguard model based on the Llama 4 architecture. It classifies user and model messages across multiple safety categories including violence, sexual content, child exploitation, and more. The model supports structured output for easy integration into LLM pipelines and is designed to help developers implement responsible AI guardrails in production applications.
Llama Guard 4 12B is a Foundation models & chat project. It focuses on detecting and categorizing unsafe or policy-violating content in LLM inputs and outputs at scale. It is built as an open-source project for developers. Llama Guard 4 12B is open source under the Open Source license. It runs on the web, the command line, and API.
Meta builds and maintains Llama Guard 4 12B, and it first shipped in 2023. The project is developed in the open on GitHub with 4.3k stars and 28 commits in the last 90 days. Among its 3 catalogued features are safety classification, policy categories, and multilingual support.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do