Model Cards / Meta AI

Llama Guard Paper

constitution7,148 words·31 min read·Aug 20, 2026·Source
Version History
Chaptered summary is still being generated for this document. Showing a heuristic brief in the meantime.
Summary
7,148-word document condensed to 98 words. Meta AI · Aug 20, 2026
TL;DR

Hakan Inan, Kartikeya Upasani, Jianfeng Chi , Rashi Rungta , Krithika Iyer, Yuning Mao, Michael Tontchev, Qing Hu , Brian Fuller, Davide Testuggine, Madian Khabsa GenAI at Meta We introduce Llama Guard, an LLM-based input-output safeguard model geared towards Human-AI conversation use cases. Our model incorporates a safety risk taxonomy, a valuable tool for categorizing a specific set of safety ri

Capability claim
  • We introduce Llama Guard, an LLM-based input-output safeguard model geared towards Human-AI conversation use cases.
Deployment scope
  • available tools use conventional transformer models that are small in size as their backbone (Markov et al., 2023; Lees et al., 2022).

Every italicized passage is a verbatim substring of the source document (checked deterministically after extraction). Field selection is heuristic — some quotes may lack surrounding context and some claims may be absent if no matching pattern appeared. For citation, open the source: original constitution · source SHA ad88eec2b0d4 · version dated Aug 20, 2026.