AI红队评估能证明什么：证据上限的量化分析

AI 摘要

本文提出AI红队评估的“证据上限”概念，即固定测试预算下评估结果能改变信念的最大倍数，并给出闭式解。研究发现，在可计算的危害率之上，适度规模的基准可严格证明模型安全性，而低于该阈值时，任何被动基准都无法提供所需的安全证据。对8个评估套件的审计显示，现有基准对高频危害有效，但对罕见灾难性危害的评估能力相差数个数量级。

正文节选

What AI Red-Team Evaluations Can and Cannot Prove Abstract Red-team evaluations of AI models support some claims and not others, and the boundary between the two is calculable rather than merely a matter of judgment. We define the evidential ceiling of an evaluation as the largest factor by which one result can move belief under a fixed testing budget, derive it in closed form for the benchmark null result, and use it to locate that boundary exactly. We find that above a calculable harm rate, a

发布时间：—

抓取时间：2026-08-07 05:58

来源机构：Hugging Face