The GGML and GGUF formats used by llama.cpp have been found to contain potentially exploitable memory corruption vulnerabilities.

•1m read time•From simonwillison.net
Post cover image
Share this post