权重全都摆开了，能力照样锁得住

有人把能力加密后塞进模型权重，四种诱导方法一个都没问出来。评估能证明的只是"诱导出了什么"，证明不了"里面没有"。 1 2

#AI 安全 #模型评估 #权重后门 #论文解读 #供应链安全

1
AI Models Can Provably Hide Arbitrary Capabilities（arXiv:2610.13928）
arxiv.org
2
Hacker News 讨论
news.ycombinator.com

This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.

Comments

1/6

2026-10-15

有人把能力加密后塞进模型权重，四种诱导方法一个都没问出来。评估能证明的只是"诱导出了什么"，证明不了"里面没有"。 1 2

#AI 安全 #模型评估 #权重后门 #论文解读 #供应链安全

1
AI Models Can Provably Hide Arbitrary Capabilities（arXiv:2610.13928）
arxiv.org
2
Hacker News 讨论
news.ycombinator.com

This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.