AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

CIMemories: A Compositional Benchmark For Contextual Integrity In LLMs

Niloofar Mireshghallah, Neal Mangaokar, Narine Kokhlikyan, Arman Zharmagambetov, Manzil Zaheer, Saeed Mahloujifar, Kamalika Chaudhuri

ICLR 2026 · Large language models › Language, safety and interpretability · approach: LLM pruning and scaling data

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase number of tasks from 1 to 40GPT-5 · CIMemoriesViolation@1: 0.1% -> 9.6%Hurt
sample 5 generations instead of 1 for the same promptGPT-5 · CIMemoriesViolation: 9.6% -> 25.1%Hurt
scale Qwen-3 model size from 1.7B to 32B (ablation)Qwen-3 family · CIMemoriesViolation@5 and completenessMixed
enable reasoning chain (ablation)Qwen-3 30B · CIMemoriesViolation@5Improved
enable reasoning chain (ablation)Qwen-3 30B · CIMemoriesCompletenessNo clear effect
add conservative privacy prompt (medium, high) (ablation)GPT-5 · CIMemoriesViolation@5Improved
add conservative privacy prompt (medium, high) (ablation)GPT-5 · CIMemoriesCompletenessHurt
add light (task-success-emphasizing) prompt (ablation)GPT-5 · CIMemoriesCompletenessImproved
increase number of inappropriate attributes in memory (ablation)GPT-5 · CIMemoriesViolations@5Hurt
increase number of inappropriate attributes in memory (ablation)GPT-5 · CIMemoriesCompletenessNo clear effect
choose GPT-4o over Qwen-3 32Bmodel choice · CIMemories (10 profiles)Violation@5 vs completeness: 14.82% vs 69.14% violation; 43.95% vs 57.63% completenessMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy