Base Labs, Hugging Face and Goodfire Launch Open-Weight AI Safety Standard as 6,000 Abliterated Models Spread
Updated
Updated · TechCrunch · Sep 17
Base Labs, Hugging Face and Goodfire Launch Open-Weight AI Safety Standard as 6,000 Abliterated Models Spread
1 articles · Updated · TechCrunch · Sep 17
Summary
Base Labs on Wednesday unveiled a new safety infrastructure standard for open-weight AI models with Hugging Face and Goodfire AI, aiming to bake evaluation and monitoring into training and deployment.
More than 6,000 abliterated models are already listed on Hugging Face, underscoring concerns that open models can be made dangerous by stripping away safeguards.
Base Labs said it will develop and publish methods for training and monitoring open models, pitching transparency and built-in controls as safer than protections added after deployment.
Technical details of the partnership were not disclosed, but Goodfire — which focuses on model interpretability — signaled the effort should embed safety in the models served.
Baseten, valued at $13 billion after a $1.5 billion Series F in June, is also inviting outside developers to help shape the framework into a broader ecosystem standard.