OpenAI 因 Astra 可能跨越为自主零日攻击设定的网络阈值，暂停了两周部署导向的 RL 训练，其最大前沿 RL 运行仍搁置。此前 Hugging Face 事件中评估模型逃逸网络边界，Astra 虽未涉及，但其评估强到 OpenAI 无法排除临界阈值。研究负载现面临更强沙箱、更严网络访问及持续安全测试，监控亦覆盖内部活动与工具操作。
OpenAI slowed frontier training because Astra may have crossed the cyber threshold built for autonomous zero-day attacks.
It paused two weeks of deployment-focused RL training, while its largest planned frontier RL run remains on hold.
This is after the Hugging Face incident, where evaluation models escaped their intended network boundary and reached production infrastructure. Astra was not involved, but its separate evaluations were strong enough that OpenAI said it could not rule out the Critical threshold.
Research workloads now face stronger sandboxing, tighter network access, fewer shared services, and continuous security testing before resuming.
Monitoring also examines internal activity and tool actions, escalating suspicious behavior to automated investigators and human reviewers.
As models become more capable, the risks associated with developing and testing them internally also grow. We temporarily paused reinforcement learning (RL) training on our latest models intended for deployment for two weeks while we hardened and red-teamed our research environments and expanded monitoring coverage. Our largest planned frontier RL run remains on hold while smaller-scale training and evaluations validate these safeguards and establish more evidence of alignment. https://openai.com/index/pacing-model-development-cyber-capabilities/在 X 查看被引用的帖子
来源：Rohan Paul · x.com