Baseten旗下研究部门Base Labs与Hugging Face和Goodfire AI达成合作，共同建立开权重AI模型的安全基础设施标准1。该伙伴关系旨在开发针对开权重模型的安全评估和监测方法1，以应对日益严重的模型安全威胁。根据统计，Hugging Face平台目前托管着超过6000个被移除安全防护的模型1。
此次合作强调将安全特性嵌入模型的训练和部署阶段1，而非采取事后补救的方式。Baseten在6月完成1.5亿美元F轮融资，公司估值达到130亿美元1。合作伙伴Goodfire AI亦在今年早期完成1.5亿美元B轮融资，由B Capital领投1。Base Labs于今年早期由Baseten成立1。
Base Labs, the research division of Baseten, has initiated a collaborative effort with Hugging Face and Goodfire AI to establish open-source infrastructure standards for AI model safety 1. The partnership aims to develop safety assessment and monitoring methodologies specifically designed for open-weight models, addressing the growing threat landscape surrounding model security 1.
The collaboration underscores a shift toward embedding safety mechanisms directly into model training and deployment processes rather than applying protective measures retroactively 1. This approach comes as Hugging Face currently hosts over 6,000 models that have had their safety guardrails removed 1. Base Labs, established by Baseten earlier this year, is spearheading this initiative to create more robust safeguards within the open-source AI ecosystem 1.
The partners bring substantial resources to the effort. Baseten completed a Series F funding round of $150 million in June, valuing the company at $13 billion 1, while Goodfire AI raised $150 million in its Series B round earlier in the year, led by B Capital 1.
评论
还没有评论，欢迎留下第一条。