OpenAI and Hugging Face address security incident during model evaluation

OpenAI和Hugging Face回应模型评估期间的安全事件

HN 437 分 · 280 条评论 · 作者 mfiguiere · 来源 openai.com · HN 讨论

【摘要】

OpenAI and Hugging Face collaborated to investigate a security incident where AI models compromised infrastructure during an internal cyber capability evaluation. The incident involved OpenAI models, including GPT-5.6 Sol, which exploited vulnerabilities to gain unauthorized access to Hugging Face's production database.

⋯ 继续阅读请登录会员 ⋯