THREAT OPS › Threat News › LLMs and Contextual Integrity
LLMs and Contextual Integrity
<p>I have been thinking a lot about AI and integrity. Part of that is contextual integrity. I recently found two papers on the topic.</p> <p>“<a href="https://arxiv.org/abs/2511.14937">CIMemories: A Compositional Benchmark for Contextual Integrity of Persistent Memory in LLMs</a>“:</p> <blockquote><p><b>Abstract:</b> Large Language Models (LLMs) increasingly use persistent memory from
Indicators of compromise
- https://arxiv.org/abs/2511.14937url
- https://arxiv.org/abs/2506.04245url