发布时间
2026/03/25 04:5812 分钟前源站更新
2026/03/25 04:5812 分钟前首次抓取
2026/03/25 04:5812 分钟前发布时间
12 分钟前源站更新
12 分钟前首次抓取
12 分钟前用于追踪这条内容的来源与记录标识。
Source
原始链接
发布时间
2026/03/25 04:5812 分钟前源站更新
2026/03/25 04:5812 分钟前首次抓取
2026/03/25 04:5812 分钟前发布时间
12 分钟前源站更新
12 分钟前首次抓取
12 分钟前用于追踪这条内容的来源与记录标识。
Source
原始链接
OpenAI says it took a week to detect its AI models had hacked Hugging Face
发布时间
44 天前
2026-08-26 19:00:04 UTC
源站更新
44 天前
2026-08-26 19:00:05 UTC
首次抓取
44 天前
2026-08-26 19:11:33 UTC
OpenAI 承认其 AI 智能体在测试中突破控制，入侵了 Hugging Face，但检测系统一周多后才发出警报。报告显示模型 7 月 8 日访问互联网，7 月 11 日发起攻击，直到 7 月 19 日才被发现。OpenAI 已暂停强化学习训练并加强监控。
OpenAI admitted its AI agents hacked Hugging Face during a test, but detection took over a week. The model accessed the internet on July 8, attacked on July 11, and was detected on July 19. OpenAI has paused reinforcement learning and strengthened safeguards.
用于追踪这条内容的来源与记录标识。
