Anthropic 向 METR 开放员工级访问权限以进行独立评估
原标题：An update on independent evaluators
AI 摘要
Anthropic 宣布已开始向非营利评估机构 METR 的六名员工提供类似内部员工的访问权限，使其能深入查看模型训练、安全防护等内部信息，并允许 METR 公开发布风险评估结果，Anthropic 无权否决但可要求对特定类别信息进行脱敏。Anthropic 同时与 SecureBio、Accenture、Faculty 等机构合作，并正在与 Hugging Face 等组织初步洽谈，还制定了三级评估者访问权限框架。此举旨在通过独立第三方评估增强前沿 AI 安全监督的公共问责。
正文节选
An update on independent evaluators Last month, we committed to stronger independent, third-party assessment of our models and safety practices. We also signed the White House Accord on Super Intelligence, in which companies pledged to partner with independent external auditors or evaluators. Today, we’re announcing that we’ve started to give employee-like access to a team from METR, a nonprofit that evaluates frontier AI models. METR will publish the results of its risk assessments. Over the pa