we had a significant security incident during evaluation of our models. we are sharing what we have learned so far. thanks to @huggingface for the partnership on this. https://t.co/2o2VfR6PIa

OpenAI 官方确认,在模型评估期间发生了一起重大安全事件,已与 Hugging Face 合作分享初步调查结果。这是 OpenAI 首次公开承认在内部模型评测环节出现安全漏洞,值得行业和用户高度关注。








