English(EN) Is Memorization Context-Sensitive? Prefix-Based Extraction Beyond Isolated Prefixes

LLM记忆具有上下文敏感性，不会被RAG消除

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-09 04:00

一项新的研究论文探讨了上下文敏感条件是否会影响大型语言模型（LLM）的记忆。该研究调查了添加指令或检索到的文档（如在检索增强生成（RAG）中所见）是否会减轻记忆，还是仅仅改变了可提取的记忆序列。研究结果表明，上下文并不会消除记忆；相反，它揭示了一个稳健可提取的记忆数据核心和一个敏感边界。虽然上下文可以抑制某些暴露并启用其他暴露，但可提取的记忆样本仍然存在，持续具有安全相关性。 AI

影响 LLM中的上下文条件不会消除记忆，即使在RAG实现中也表明存在持续的安全风险。

排序理由关于LLM行为的学术论文。[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 →

本文如何被排名

Signal score: 20 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 关于LLM行为的学术论文。[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.AI TIER_1 English(EN) · Ali Satvaty, Narjes Sharafi, Jirui Qi, Suzan Verberne, Fatih Turkmen · 2026-10-09 04:00

记忆是否具有上下文敏感性？超越孤立前缀的基于前缀的提取

arXiv:2610.12085v1 Announce Type: new Abstract: Large language models (LLMs) can expose memorized training sequences under prefix-based extraction: given a prefix from a training example, the model may assign high probability to the original continuation. In deployed systems, how…

本文如何被排名

报道来源 [1]

记忆是否具有上下文敏感性？超越孤立前缀的基于前缀的提取

相关实体

相关话题