1
LLMs and Contextual Integrity
大模型持久记忆暗藏隐私雷区，用情境完整性框架守住信息边界，AI安全必修课。
I have been thinking a lot about AI and integrity. Part of that is contextual integrity. I recently found two papers on the topic. “ CIMemories:…
2
Plotline – a context-integrity benchmark for LLM apps, and the fixes it drove
一个专测LLM上下文完整性的基准，还附带了实测驱动的修复方案，做AI应用避坑必备。
Article URL: https://github.com/Doric-builder/plotline Comments URL: https://news.ycombinator.com/item?id=48814603 Points: 1 # Comments: 0
3
It Takes Two: Complementary Self-Distillation for Contextual Integrity in LLMs
互补自蒸馏如何维护大模型上下文完整性？这项研究提出双模型协作新方案，为LLM安全对齐提供创新思路。
arXiv:2605.20258v1 Announce Type: new Abstract: Contextual Integrity (CI) defines privacy not merely as keeping information hidden, but as governing i…