Baseten's Base Labs Teams With Hugging Face and Goodfire on Open-Weight AI Safety Standard

Baseten introduced a safety infrastructure standard through its Base Labs research arm on Wednesday, working with Hugging Face and Goodfire AI to create evaluation and monitoring tools for open-weight models.

The move comes as open-weight models face scrutiny over a technique called abliteration, which strips away safeguards. Hugging Face, which hosts open source AI models, lists more than 6,000 abliterated models.

Base Labs, launched by Baseten earlier this year, will develop and publish training and monitoring methods. The company describes its aim as a transparent "standard" for open models, integrated into training and deployment rather than added afterward.

"We believe openness to be an advantage for AI safety," the company said on X, adding that openness offers more visibility into model behavior and better ways to turn safety research into controls than closed-source approaches.

Technical details of the partnership were not disclosed. Goodfire, which works on explaining how models make decisions, said in a reply to Baseten's post that "Safety must be built into open models and provided by those who serve them."

Baseten, an AI inference provider, raised a $1.5 billion Series F in June, reaching a $13 billion valuation. Goodfire raised a $150 million Series B led by B Capital earlier this year for its model interpretability platform.

Baseten is inviting the wider developer ecosystem to contribute to the framework, saying, "Together, we are building an ecosystem of open models that are safe and accessible to all."

Source: TechCrunch