We believe that AI models benefit from an open approach, both in terms of innovation and safety. Releasing models like Code Llama means the entire community can evaluate their capabilities, identify issues & fix vulnerabilities.

→ View original post on X — @aiatmeta,