mlsanityData to 14 Oct 2026

1-2-3 Check: Enhancing Contextual Privacy in LLM via Multi-Agent Reasoning

arXiv 2508.07667 ↗First posted 11 Aug 2025cs.AIcited by 17

On the map: LLM reasoning → Reasoning for agents, planning and code

also in 1 more
Abstract (arXiv)

Addressing contextual privacy concerns remains challenging in interactive settings where large language models (LLMs) process information from multiple sources (e.g., summarizing meetings with private and public information). We introduce a multi-agent framework that decomposes privacy reasoning into specialized subtasks (extraction, classification), reducing the information load on any single agent while enabling iterative validation and more reliable adherence to contextual privacy norms. To understand how privacy errors emerge and propagate, we conduct a systematic ablation over information-flow topologies, revealing when and why upstream detection mistakes cascade into downstream leakage. Experiments on the ConfAIde and PrivacyLens benchmark with several open-source and closed-sourced LLMs demonstrate that our best multi-agent configuration substantially reduces private information leakage (18% on ConfAIde and 19% on PrivacyLens with GPT-4o) while preserving the fidelity of public content, outperforming single-agent baselines. These results highlight the promise of principled information-flow design in multi-agent systems for contextual privacy with LLMs.

Key facts. Cited by 17 papers in the map (data through 14 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 30 of its 43 references that are in the map. First posted on arXiv 11 Aug 2025.

Builds on

30 of 43 references.

References · 30

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2024The Llama 3 Herd of Modelscited by 14,715
  3. 2024GPT-4o System Cardcited by 4,588
  4. 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,332
  5. 2022Quantifying Memorization Across Neural Language Modelscited by 763
  6. 2024WildChat: 1M ChatGPT Interaction Logs in the Wildcited by 629
  7. 2023Multi-Agent Collaboration: Harnessing the Power of Intelligent LLM Agentscited by 333
  8. 2023Can LLMs Keep a Secret? Testing Privacy Implications of Language Models via Contextual Integrity Theorycited by 230
  9. 2022What Does it Mean for a Language Model to Preserve Privacy?cited by 195
  10. 2024PrivacyLens: Evaluating Privacy Norm Awareness of Language Models in Actioncited by 171
  11. 2020Moral Stories: Situated Reasoning about Norms, Intents, Actions, and their Consequencescited by 135
  12. 2024AirGapAgent: Protecting Privacy-Conscious Conversational Agentscited by 94
18 more
  1. 2023Moral Foundations of Large Language Modelscited by 85
  2. 2024Follow My Instruction and Spill the Beans: Scalable Data Extraction from Retrieval-Augmented Generation Systemscited by 76
  3. 2021What Would Jiminy Cricket Do? Towards Agents That Behave Morallycited by 57
  4. 2023Privacy in Large Language Models: Attacks, Defenses and Future Directionscited by 51
  5. 2024MegaAgent: A Large-Scale Autonomous LLM-based Multi-Agent System Without Predefined SOPscited by 47
  6. 2024GoldCoin: Grounding Large Language Models in Privacy Laws via Contextual Integrity Theorycited by 37
  7. 2024Operationalizing Contextual Integrity in Privacy-Conscious Assistantscited by 37
  8. 2024CI-Bench: Benchmarking Contextual Integrity of AI Assistants on Synthetic Datacited by 37
  9. 2024LLMArena: Assessing Capabilities of Large Language Models in Dynamic Multi-Agent Environmentscited by 36
  10. 2024CMAT: A Multi-Agent Collaboration Tuning Framework for Enhancing Small Language Modelscited by 26
  11. 2025Protecting Users From Themselves: Safeguarding Contextual Privacy in Interactions with Conversational Agentscited by 26
  12. 2024Unique Security and Privacy Threats of Large Language Models: A Comprehensive Surveycited by 23
  13. 2024Perceptions to Beliefs: Exploring Precursory Inferences for Theory of Mind in Large Language Modelscited by 22
  14. 2024Privacy Checklist: Privacy Violation Detection Grounding on Contextual Integrity Theorycited by 18
  15. 2023Are Chatbots Ready for Privacy-Sensitive Applications? An Investigation into Input Regurgitation and Prompt-Induced Sanitizationcited by 17
  16. 2025Position: Contextual Integrity is Inadequately Applied to Language Modelscited by 11
  17. 2024Can LLMs get help from other LLMs without revealing private information?cited by 9
  18. 2024NAT-NL2GQL: A Novel Multi-Agent Framework for Translating Natural Language to Graph Query Languagecited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 3; 2026: 14.

17 papers.

Sort:

By area:

  1. 2025MAGPIE: A benchmark for Multi-AGent contextual PrIvacy Evaluationcited by 29
  2. 2025Position: Privacy Is Not Just Memorization!cited by 17
  3. 2026Information-Theoretic Privacy Control for Sequential Multi-Agent LLM Systemscited by 8
  4. 2026PrivAct: Internalizing Contextual Privacy Preservation via Multi-Agent Preference Trainingcited by 7
  5. 2026Contextualized Privacy Defense for LLM Agentscited by 7
  6. 2026AgentRaft: Automated Detection of Data Over-Exposure in LLM Agentscited by 6
  7. 2026POLAR-Bench: A Diagnostic Benchmark for Privacy-Utility Trade-offs in LLM Agentscited by 4
  8. 2026PrivacyReasoner: Can LLM Emulate a Human-like Privacy Mind?cited by 2
  9. 2026Do LLMs Know What Is Private Internally? Probing and Steering Contextual Privacy Norms in Large Language Model Representationscited by 2
  10. 2026PrivacyAlign: Contextual Privacy Alignment for LLM Agentscited by 2
  11. 2026PiSAs: Benchmarking Contextual Integrity in Multi-User Agentic Systemscited by 2
  12. 2025User Perceptions vs. Proxy LLM Judges: Privacy and Helpfulness in LLM Responses to Privacy-Sensitive Scenarioscited by 1
5 more
  1. 2026It Takes Two: Complementary Self-Distillation for Contextual Integrity in LLMscited by 1
  2. 2026SoK: When Safe Agents Fail Together: The Security of Multi Agent LLM Systemscited by 1
  3. 2026Agents That Know Too Much: A Data-Centric Survey of Privacy in LLM Agentscited by 0
  4. 2026Value-Preserving Architectures for Agentic AI Systemscited by 0
  5. 2026When Agents Talk, Secrets Travel: Topology-Driven Leakage and Recipient-Conditioned Sanitization in Multi-Agent LLM Systemscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025MAGPIE: A dataset for Multi-AGent contextual PrIvacy Evaluation
  2. 2026AgentLeak: A Benchmark for Internal-Channel Privacy Leakage in Multi-Agent LLM Systems
  3. 2026NeuroFilter: Activation-Based Guardrails for Privacy-Conscious LLM Agents
  4. 2026Privacy-Preserving RAG via Multi-Agent Semantic Rewriting: Achieving Confidentiality Without Compromising Contextual Fidelity
  5. 2025Privacy in Action: Towards Realistic Privacy Mitigation and Evaluation for LLM-Powered Agents
  6. 2025Privacy Reasoning in Ambiguous Contexts
  7. 2026OCELOT: Inference-Leakage Budgets for Privacy-Preserving LLM Agents
  8. 2025Personalizing Agent Privacy Decisions via Logical Entailment