Base Labs partners with Hugging Face on open-weight AI safety

Baseten has launched Base Labs, a new research initiative focused on AI safety for open-weight models, in partnership with Hugging Face and Goodfire AI. The collaboration aims to establish a transparency standard for monitoring and training open models, addressing the rising issue of abliteration techniques that remove model safeguards. Hugging Face currently hosts over 6,000 such compromised models, highlighting the scale of the problem. While technical details remain undisclosed, the partnership leverages Goodfire’s interpretability platform to embed safety directly into model development. Baseten is inviting broader developer contributions to build this ecosystem.