Privacy-Preserving Language Models

Ask about this topic

Overview

This field studies how to use and adapt language models without exposing private training data, user prompts, or personal context. Approaches include differential privacy for fine-tuning and preference alignment, federated and split learning, cryptographic inference, and practical safeguards such as on-device processing, prompt sanitization, and selective unlearning. Researchers evaluate both formal guarantees and empirical attacks, including membership inference, data extraction, and reconstruction from intermediate activations, across language tasks and clinical or conversational settings. The central tension is preserving privacy without sacrificing model quality, cost, or speed, while ensuring evaluations reflect realistic, multi-turn use and adaptive attackers.

Latest papers

CardsList

178 papers26 papers in the last four weeks, up 8% on the four weeks before. 0.2% of all new papers.

  1. Answer-Only Disclosure: Private-Context LLM Assistance over Web APIs with a Leakage Ledger

    Oct 28, 2026Liang Yuan, Xiaohu Du, Ruikun Luo +6Privacy Leakage in Language ModelsPrivacy-Preserving Language Models

    Australian Catholic University North Sydney, NSW, Australia · Ant Group China · University of Macau Macau, China +5

  2. Reversible Pseudonymisation of Documents Sent to External LLMs: an Open Bilingual Benchmark and an Engine

    Oct 25, 2026Vladimir BeskorovainyiPrivacy Leakage in Language ModelsPII Detection

  3. EdgeDeID: A 0.6B Clinical De-identifier Trained Without Real Patient Data

    Oct 24, 2026Sina Daneshgar, Raya Namazi, Amirali Khosravi +2Privacy-Preserving Language Models

    Sharif University of Technology · Department of Radiology, Mayo Clinic · Department of Radiology and Biomedical Imaging, Yale University

  4. Privacy in Context Is Being Overlooked in the Age of General-Purpose LLMs

    Oct 21, 2026Prakhar Ganesh, Yash More, Marco Romanelli +2Privacy-Preserving Language Models

    McGill University · Mila - Quebec AI Institute · Hofstra University +1

  5. Contextual Candour: Role-Aware Graduated Disclosure for LLMs over Mixed-Sensitivity Contexts

    Oct 20, 2026Devansh Srivastav, David Pape, Samira Abedini +1LLM SecurityPrivacy Leakage in Language Models

    CISPA Helmholtz Center for Information Security, Saarbrücken, Germany

  6. Privacy-preserving inference on hosted models: a learned input-embedding transform for frozen Qwen3.8-27B

    Oct 20, 2026Mahendra Roopa, Srikanth RoopaPrivacy-Preserving Language ModelsLLM Inference

    Spinal

  7. Profile-DP: Differential Privacy for User-Controlled Profile Release in Personalized LLM Inference

    Oct 19, 2026Mingchen Li, Zifan Peng, Kewei Sha +5Privacy-Preserving Language ModelsDifferential Privacy

    University of North Texas · School of Computing, Newcastle University, UK · Augusta University

  8. Veil-Then-Reveal: Protecting PII Privacy via Reversible Pseudo-Anonymization in Zero-Trust User--LLM Interaction

    Oct 16, 2026Mohamed Elmahallawy, Mohamed Shaaban, Thomas E. CarrollPrivacy-Preserving Language Models

    School of Engineering and Applied Sciences, Washington State University, Richland, WA 99354, USA · Pacific Northwest National Laboratory, Richland, WA 99354, USA

  9. Prune the Leaky Neurons: Mitigating Private Data Leakage in Fine-Tuned LLMs via Joint-Embedding-Guided Pruning

    Oct 16, 2026Mohamed Shaaban, Mohamed ElmahallawyPrivacy Leakage in Language ModelsPrivacy-Preserving Language Models

    School of Engineering and Applied Sciences, Washington State University, Richland, WA 99354, USA

  10. Exploring The Empirical Privacy Landscape of Synthetic Data

    Oct 15, 2026Zihao Zhao, Avi Schwarzschild, Casey Meehan +3Privacy Leakage in Language ModelsPrivacy-Preserving Language Models

    Johns Hopkins University · OpenAI

  11. FedSkillGit: Federated Skill Evolution via Version Control

    Oct 15, 2026Boxun Lan, Jiamei Li, Weijian Li +5LLM Agent Skill LearningFederated Learning

    East China Normal University, Shanghai, China · Shandong University, Jinan, China

  12. PocketTSLM: Deploying an On-Device Soft-Prompt Time-Series Language Model for Privacy-Preserving Smartphone Reasoning

    Oct 15, 2026Max Rosenblattl, Nikhil Krishnaswamy, Leon Nissen +9Time Series Language ModelingOn-Device Language Model Inference

    Stanford University, USA · CHI, Technical University of Munich University Hospital, Germany · ETH Zurich, Switzerland +3

  13. A Graph-Based Analysis of Comparability, Scale, and Attribution in Private Transformer Inference

    Oct 14, 2026Souvik Das, Ariful Azad, Nitesh SaxenaTransformer InferencePrivacy-Preserving Language Models

    Texas A&M University, College Station, TX, USA

  14. Cost-Efficient Privacy Enforcement for Email-Generating LLM Agents under Strategic Information Requests

    Oct 13, 2026Qikun Cai, Tao Huang, Chen HouPrivacy Protection for LLM AgentsLLM Agent Security

  15. Differential Privacy Protected Preference Alignment on Survey Derived Data

    Oct 12, 2026Xiuyuan Qi, Yi Wang, Wenjie WangPrivacy-Preserving Language ModelsPreference Alignment

    ShanghaiTech University

  16. DP-MIND: Differentially private memory-guided Long context Inference

    Oct 9, 2026Zhihao Liu, Yitian Cheng, Shuang Liu +3Privacy-Preserving Language ModelsLong-Context Language Model Inference

    Zhejiang University · Ant Group · Institute of Science Tokyo +1

  17. Sensitive-Topic Leakage Through LLM Routing Metadata: Measurement and Mitigation

    Oct 7, 2026Teng-Ruei ChenPrivacy Leakage in Language ModelsLLM Routing

    Krixvon, Taipei 100, Taiwan

  18. zkLLMPoT: Efficient Zero Knowledge Proof of Training for Large Language Models

    Oct 6, 2026Junkai Liang, Zhanpeng Guo, Pengfei Wu +5Zero-Knowledge MLLLM Auditing

    Peking University · Singapore Management University · China Telecom Quantum Information Technology Group Co., Ltd +1

  19. Reward-Driven Learning under Prompt-Level Differential Privacy

    Oct 5, 2026Jiachen Zhao, Antonia Januszewicz, Taeho JungPrivacy-Preserving Language ModelsDifferentially Private Stochastic Gradient Descent

    University of Notre Dame

  20. Static Bootstrap Placement for Encrypted Language Model Decoding

    Oct 4, 2026Halil Ibrahim Kanpak, Didem UnatPrivacy-Preserving Language ModelsLLM Inference

    Department of Computer Engineering, Koç University, Istanbul, Türkiye

  21. Inner Momentum for Differentially Private Muon

    Oct 2, 2026Bishnu Bhusal, Minh Vu, Ben Southworth +3Muon OptimizerPrivacy-Preserving Language Models

    Los Alamos National Laboratory · University of Missouri · Georgia Institute of Technology

  22. Mitigating Private Data Leakage in LLMs with Whiteout

    Oct 1, 2026Anna Yoo Jeong Ha, Ronik Bhaskar, Haitao Zheng +1Privacy Leakage in Language ModelsPrivacy-Preserving Language Models

    Department of Computer Science, University of Chicago

  23. SIEVE: Selective attention-value Suppression for Vision-Language Models Unlearning

    Oct 1, 2026Si Qi Goh, Cap Dang Xuan Kiet, Tat-Jen Cham +1VLM UnlearningPrivacy-Preserving Language Models

    Digital Trust Centre, Nanyang Technological University, Singapore · College of Computing and Data Science, Nanyang Technological University, Singapore

  24. Is Weight Tying Still Beneficial for Decoder-Only LLMs in Private Settings Under DP-SGD?

    Sep 30, 2026Razan El Mais, Ali Chehab, Ibrahim Issa +1Privacy-Preserving Language ModelsDifferentially Private Stochastic Gradient Descent

    Department of Electrical and Computer Engineering American University of Beirut Beirut, Lebanon

  25. PrivMeSA: Privacy-Aware Self-Evolving Multi-Agent System for Medicine via Local-Remote LLM Collaboration

    Sep 29, 2026Dannong Wang, Yuran Zhang, Bian Sun +4Multi-Agent LLM SystemsPrivacy-Preserving Language Models

    Institute of Artificial Intelligence, University of Central Florida · Department of Computer Science and Operations Research, Universit´e de Montr´eal · Department of Medicine, University of Central Florida

  26. Tokenized Key-Gated Adapter Routing: A Secure Access Control Mechanism Against Private Data Leakage in LLMs

    Sep 28, 2026Mohamed Shaaban, Mohamed ElmahallawyLLM SecurityPrivacy Leakage in Language Models

    School of Engineering and Applied Sciences, Washington State University, Richland, WA 99354, USA

  27. "Nothing to See Here'': Unintended Disclosure through Revision Traces of LLM Deliverables

    Sep 28, 2026Yage Zhang, Yukun Jiang, Yang ZhangPrivacy-Preserving Language ModelsLLM Auditing

    CISPA Helmholtz Center for Information Security

  28. CipherGenome: Homomorphic Inference for Genomic Mixture-of-Experts

    Sep 27, 2026Guang Yang, Fengchen LiuMixture-of-Experts InferencePrivacy-Preserving Language Models

    University of California, Los Angeles · University of California, Berkeley

  29. GenomeOcean Anywhere: Private WebGPU Inference for Genome MoEs

    Sep 27, 2026Guang Yang, Fengchen LiuGenomic Foundation ModelsDistributed Inference

    University of California, Los Angeles · University of California, Berkeley

  30. Learning to Refer: Client-Resolved Generation for Privacy-Aware Language Models

    Sep 26, 2026Jeongho Yoon, Chanhee Park, Yongchan Chun +5Privacy-Preserving Language ModelsPrivacy Leakage in Language Models

    Department of Computer Science and Engineering, Korea University · Konkuk University · Soongsil University +1