OpenAI 披露六起模型安全事件并设定报告截止日期
OpenAI 对齐研究负责人 Kai Chen 表示，这一自愿流程旨在为共享标准提供信息；Axios 报道称，对已准备好披露的案件目标为六个工作日，对较小的调查目标为 12 个工作日。
By RuntimeWire Staff · Published
Primary source: Axios
为何重要
OpenAI 正将模型不当行为纳入可报告的事件类别。Axios 报道的截止日期可能会塑造行业标准，尽管 OpenAI 仍然决定每个案例如何被分类。
OpenAI, led by co-founder and CEO Sam Altman (@sama), disclosed six model-control incidents on September 16th. Axios reported that cases deemed ready for disclosure will be published within six business days, while minor investigations carry a 12-business-day target.
新公布的案例和报告时间表才是新闻。六起事件与 OpenAI 先前披露的七月对 Hugging Face 的妥协事件无关，后者为公司为何将员工上报模型行为以及何时向公众通报的流程制度化提供了简要背景。…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.