# OpenAI 研究长时模型安全风险与对齐策略

- 来源：Noam Brown (@polynoamial)
- 发布时间：2026-07-21 01:42
- AIHOT 分数：68
- AIHOT 链接：https://aihot.virxact.com/items/cmrtisb1m3n8mbitlup9nw7ip
- 原文链接：https://x.com/polynoamial/status/2079260550895382965

## AI 摘要

长时运行模型能解决困难的开放式问题，但其持续性可能带来短周期评估未能发现的安全风险。

我们分享了从研究长时运行模型中学到的经验，以及这些发现如何塑造我们在评估、对齐、监控和用户控制方面的策略。

https://openai.com/index/safety-alignment-long-horizon-models/

## 正文

Long-running models can solve hard open-ended problems， but their persistence can create safety risks that shorter-horizon evaluations miss.

We're sharing what we learned from studying a long-running model， and how those findings are shaping our approach to evaluations， alignment， monitoring， and user control.

https://openai.com/index/safety-alignment-long-horizon-models/