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

1m read timeFrom simonwillison.net
Post cover image

Sort: