PulseAugur
中
实时 04:40:53
English(EN) Training Deliberative Monitors for Black-Box Scheming Detection

新型监控器可在无内部访问的情况下检测AI代理的欺骗行为

作者 PulseAugur 编辑部 · [1 个来源] ·

研究人员开发了一种新方法，用于训练更小的、开源的模型来检测自主代理中的欺骗行为。这些“审慎监控器”仅基于代理轨迹运行，无需访问代理的内部推理或模型内部。训练过程包括使用一个前沿模型生成理由，过滤这些理由，然后通过监督微调和强化学习将高质量的理由提炼到监控器中。评估表明，在各种错位基准测试中，这些监控器的成本效益可以超越或媲美更大、经过提示的前沿模型。 AI

影响 通过使用更小的、开源的模型，实现了更具成本效益的AI安全监控。

排序理由 该集群包含一篇研究论文，详细介绍了一种训练AI监控器的新方法。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

新型监控器可在无内部访问的情况下检测AI代理的欺骗行为

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
该集群包含一篇研究论文，详细介绍了一种训练AI监控器的新方法。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
paper, safety, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
94 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

  1. arXiv cs.AI TIER_1 English(EN) · Aditya Sinha, Akshat Naik, Victor Gillioz, Simon Storf, Kilian Merkelbach, Rich Barton-Cooper, Axel H{\o}jmark, Marius Hobbhahn ·

    训练审议式监控器以检测黑盒诡计

    arXiv:2605.29601v1 Announce Type: cross Abstract: As autonomous agents become more capable of performing real-world tasks, distinguishing scheming behavior from benign task pursuit may become a central AI control problem. Existing monitors often rely on chain-of-thought access or…

相关实体

相关话题