English(EN) Extracting Forgotten Prompts from Targeted Unlearned Models

新型攻击可从未学习的AI模型中提取遗忘的提示

作者 PulseAugur 编辑部 · [1 个来源] · 2026-09-04 04:00

研究人员开发了一种名为目标主动搜索（TAS）的新型攻击，可以从未学习的AI模型中提取遗忘的提示。与先前假设已知遗忘提示的方法不同，TAS利用保留的数据和黑盒访问来识别和重构提示本身。实验表明，TAS在识别遗忘实体方面达到了100%的准确率，并且与朴素探测方法相比，查询次数显著减少的情况下，重构了高达95%的提示。 AI

影响这项研究突显了AI遗忘技术的一个潜在漏洞，表明即使在数据删除后，敏感信息仍可能被恢复。

排序理由该集群包含一篇学术论文，详细介绍了一种从AI模型中提取信息的新方法。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

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

本文如何被排名

Signal score: 15 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇学术论文，详细介绍了一种从AI模型中提取信息的新方法。[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; 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: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.LG TIER_1 English(EN) · Au Ashley Hoi-Ting, Meghdad Kurmanji, William F. Shen, Nicholas D. Lane, Ligang He · 2026-09-04 04:00

从目标未学习模型中提取遗忘提示

arXiv:2609.03662v1 Announce Type: new Abstract: Recent unlearning methods (e.g. NPO, DPO, LUNAR) make use of refusal alignment to suppress forgotten data. However, it has been shown that refusal responses might leave traces of unlearning, and recent attacks have been able to succ…

本文如何被排名

报道来源 [1]

从目标未学习模型中提取遗忘提示

相关实体

相关话题