CoChat

LLM-PBE: Assessing Data Privacy in Large Language Models

Open access

The DOI matches its Crossref record; 94 of the 94 references it cites were found in a registry or an index.

Abstract

Large Language Models (LLMs) have become integral to numerous domains, significantly advancing applications in data management, mining, and analysis. Their profound capabilities in processing and interpreting complex language data, however, bring to light pressing concerns regarding data privacy, especially the risk of unintentional training data leakage. Despite the critical nature of this issue, there has been no existing literature to offer a comprehensive assessment of data privacy risks in LLMs. Addressing this gap, our paper introduces LLM-PBE, a toolkit crafted specifically for the systematic evaluation of data privacy risks in LLMs. LLM-PBE is designed to analyze privacy across the entire lifecycle of LLMs, incorporating diverse attack and defense strategies, and handling various data types and metrics. Through detailed experimentation with multiple LLMs, LLM-PBE facilitates an in-depth exploration of data privacy concerns, shedding light on influential factors such as model size, data characteristics, and evolving temporal dimensions. This study not only enriches the understanding of privacy issues in LLMs but also serves as a vital resource for future research in the field. Aimed at enhancing the breadth of knowledge in this area, the findings, resources, and our full technical report are made available at https://llm-pbe.github.io/, providing an open platform for academic and practical advancements in LLM privacy assessment.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.14778/3681954.3681994

Cites

94 of 94 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Research works printed but not linked: 41

  1. [18] Angela Bonifati, Sihem Amer-Yahia, Chen Lei, Li Guoliang, Shim Kyuseok, Xu Jianliang, and Yang Xiaochun. 2023. From Large Language Models to Databases and Back A discussion on research and education. SIGMOD record (2023). -
  2. [22] Zhiqi Bu, Yu-Xiang Wang, Sheng Zha, and George Karypis. 2023. Automatic clipping: Differentially private deep learning made easier and stronger. Conference on Neural Information Processing Systems (2023). -
  3. [23] Zhiqi Bu, Yu-Xiang Wang, Sheng Zha, and George Karypis. 2023. Differentially private bias-term only fine-tuning of foundation models. In The Eleventh International Conference on Learning Representations . -
  4. [24] Zhiqi Bu, Yu-Xiang Wang, Sheng Zha, and George Karypis. 2023. Differentially private optimization on large model at small cost. In International Conference on Machine Learning . PMLR, 3192-3218. -
  5. [27] Nicholas Carlini, Daphne Ippolito, Matthew Jagielski, Katherine Lee, Florian Tramer, and Chiyuan Zhang. 2023. Quantifying Memorization Across Neural Language Models. In The Eleventh International Conference on Learning Representations . https://openreview.net/forum?id=TatRHT\_1cK Accessed: 2024-07-16. -
  6. [28] Nicholas Carlini, Chang Liu, Úlfar Erlingsson, Jernej Kos, and Dawn Song. 2019. The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networks. In 28th USENIX Security Symposium, USENIX Security 2019 . -
  7. [30] Nicholas Carlini, Florian Tramer, Eric Wallace, Matthew Jagielski, Ariel HerbertVoss, Katherine Lee, Adam Roberts, Tom Brown, Dawn Song, Ulfar Erlingsson, Alina Oprea, and Colin Raffel. 2021. Extracting training data from large language models. In 30th USENIX Security Symposium (USENIX Security 21) . 2633-2650. -
  8. [31] Ilias Chalkidis, Ion Androutsopoulos, and Nikolaos Aletras. 2019. Neural legal judgment prediction in English. arXiv preprint arXiv:1906.02059 (2019). - · arXiv
  9. [35] Alfredo Cuzzocrea and Hossain Shahriar. 2017. Data masking techniques for NoSQL database security: A systematic review. In 2017 IEEE International Conference on Big Data (Big Data) . IEEE, 4467-4473. -
  10. [39] Cynthia Dwork. 2006. Differential privacy. In International colloquium on automata, languages, and programming . Springer, 1-12. -
  11. [42] Vitaly Feldman and Chiyuan Zhang. 2020. What neural networks memorize and why: Discovering the long tail via influence estimation. Advances in Neural Information Processing Systems 33 (2020), 2881-2891. -
  12. [47] Tanishq Gupta, Mohd Zaki, NM Anoop Krishnan, and Mausam. 2022. MatSciBERT: A materials domain language model for text mining and information extraction. npj Computational Materials 8, 1 (2022), 102. -
  13. [51] Sorami Hisamoto, Matt Post, and Kevin Duh. 2020. Membership inference attacks on sequence-to-sequence models: Is my data in your machine translation system? Transactions of the Association for Computational Linguistics 8 (2020), 49-63. -
  14. [52] Junyuan Hong, Jiachen T Wang, Chenhui Zhang, LI Zhangheng, Bo Li, and Zhangyang Wang. 2024. DP-OPT: Make Large Language Model Your PrivacyPreserving Prompt Engineer. In The Twelfth International Conference on Learning Representations . -
  15. [54] Jie Huang, Hanyin Shao, and Kevin Chen-Chuan Chang. 2022. Are Large PreTrained Language Models Leaking Your Personal Information? EMNLP Findings (2022). -
  16. [60] Nikhil Kandpal, Eric Wallace, and Colin Raffel. 2022. Deduplicating training data mitigates privacy risks in language models. In International Conference on Machine Learning . PMLR, 10697-10707. -
  17. [65] Yannis Kotidis and Nick Roussopoulos. 1999. Dynamat: A dynamic view management system for data warehouses. ACMSigmod record 28, 2 (1999), 371-382. -
  18. [66] Yannis Kotidis and Nick Roussopoulos. 2001. A case for dynamic view management. ACM Transactions on Database Systems (TODS) 26, 4 (2001), 388-423. -
  19. [67] Vinayshekhar Bannihatti Kumar, Rashmi Gangadharaiah, and Dan Roth. 2023. Privacy Adhering Machine Un-learning in NLP. In Findings of the Association for Computational Linguistics: IJCNLP-AACL 2023 (Findings) . 268-277. -
  20. [68] Eric Lehman, Sarthak Jain, Karl Pichotta, Yoav Goldberg, and Byron C Wallace. 2021. Does BERT pretrained on clinical notes reveal sensitive data? arXiv preprint arXiv:2104.07762 (2021). - · arXiv
  21. [70] Xuechen Li, Daogao Liu, Tatsunori B Hashimoto, Huseyin A Inan, Janardhan Kulkarni, Yin-Tat Lee, and Abhradeep Guha Thakurta. 2022. When Does Differentially Private Learning Not Suffer in High Dimensions? Advances in Neural Information Processing Systems 35 (2022), 28616-28630. -
  22. [71] Xuechen Li, Florian Tramer, Percy Liang, and Tatsunori Hashimoto. 2022. Large language models can be strong differentially private learners. The Tenth International Conference on Learning Representations (2022). -
  23. [72] Yansong Li, Zhixing Tan, and Yang Liu. 2023. Privacy-preserving prompt tuning for large language model services. arXiv preprint arXiv:2305.06212 (2023). - · arXiv
  24. [75] Nils Lukas, Ahmed Salem, Robert Sim, Shruti Tople, Lukas Wutschitz, and Santiago Zanella-Béguelin. 2023. Analyzing leakage of personally identifiable information in language models. IEEE Symposium on Security and Privacy (S&P) (2023). -
  25. [77] Justus Mattern, Zhijing Jin, Benjamin Weggenmann, Bernhard Schoelkopf, and Mrinmaya Sachan. 2022. Differentially Private Language Models for Secure Data Sharing. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing . Association for Computational Linguistics, Abu Dhabi, United Arab Emirates, 4860-4873. https://aclanthology.org/2022.emnlpmain.323 Accessed: 2024-07-16. -
  26. [79] Natalie Maus, Patrick Chao, Eric Wong, and Jacob R Gardner. 2023. Black box adversarial prompting for foundation models. In The Second Workshop on New Frontiers in Adversarial Machine Learning . -
  27. [80] Luca Melis, Congzheng Song, Emiliano De Cristofaro, and Vitaly Shmatikov. 2019. Exploiting unintended feature leakage in collaborative learning. In 2019 IEEE symposium on security and privacy (SP) . IEEE, 691-706. -
  28. [86] Sharan Narang and Aakanksha Chowdhery. 2022. Pathways language model (palm): Scaling to 540 billion parameters for breakthrough performance. Google AI Blog (2022). -
  29. [88] Milad Nasr, Shuang Songi, Abhradeep Thakurta, Nicolas Papernot, and Nicholas Carlin. 2021. Adversary instantiation: Lower bounds for differentially private machine learning. In 2021 IEEE Symposium on security and privacy (SP) . IEEE, 866-882. -
  30. [92] Ashwinee Panda, Christopher A. Choquette-Choo, Zhengming Zhang, Yaoqing Yang, and Prateek Mittal. 2024. Teach LLMs to Phish: Stealing Private Information from Language Models. In The Twelfth International Conference on Learning Representations . https://openreview.net/forum?id=qo21ZlfNu6 Accessed: 2024-07-16. -
  31. Ashwinee Panda, Zhengming Zhang, Yaoqing Yang, and Prateek Mittal. 2023. Teach GPT To Phish. In The Second Workshop on New Frontiers in Adversarial Machine Learning. https://openreview.net/forum?id=tGvWCD9BEP Accessed: 2024-07-16.
  32. [98] Ricardo Jorge Santos, Jorge Bernardino, and Marco Vieira. 2011. A data masking technique for data warehouses. In Proceedings of the 15th Symposium on International Database Engineering & Applications . 61-69. -
  33. [103] Robin Staab, Mark Vero, Mislav Balunovic, and Martin Vechev. 2024. Beyond Memorization: Violating Privacy via Inference with Large Language Models. In The Twelfth International Conference on Learning Representations . https: //openreview.net/forum?id=kmn0BhQk7p Accessed: 2024-07-16. -
  34. [110] Immanuel Trummer. 2023. From bert to gpt-3 codex: harnessing the potential of very large language models for data management. arXiv preprint arXiv:2306.09339 (2023). - · arXiv
  35. [113] Thomas Vakili and Hercules Dalianis. 2021. Are Clinical BERT Models Privacy Preserving? The Difficulty of Extracting Patient-Condition Associations.. In HUMAN@ AAAI Fall Symposium . -
  36. [114] Yuli Vasiliev. 2020. Natural language processing with Python and spaCy: A practical introduction . No Starch Press. -
  37. [115] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Advances in neural information processing systems 30 (2017). -
  38. [130] Xiang Yue, Huseyin A Inan, Xuechen Li, Girish Kumar, Julia McAnallen, Huan Sun, David Levitan, and Robert Sim. 2023. Synthetic text generation with differential privacy: A simple and practical recipe. ACL (2023). -
  39. [136] Zijie Zhang, Yang Zhou, Xin Zhao, Tianshi Che, and Lingjuan Lyu. 2022. Prompt certified machine unlearning with randomized gradient smoothing and quantization. Advances in Neural Information Processing Systems 35 (2022), 1343313455. -
  40. [137] Xuandong Zhao, Lei Li, and Yu-Xiang Wang. 2022. Provably Confidential Language Modelling. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies . 943-955. -
  41. [138] Xuanhe Zhou, Zhaoyan Sun, and Guoliang Li. 2024. DB-GPT: Large Language Model Meets Database. Data Science and Engineering (2024), 1-10. -

Cited by

See all 153

Metadata

Authors
Li, Q., Hong, J., Xie, C., Tan, J., Xin, R., Hou, J., Yin, X., Wang, Z., Hendrycks, D., Wang, Z., Li, B., He, B., & Song, D.
Affiliations
  • University of California, Berkeley (US)
  • The University of Texas at Austin (US)
  • University of Illinois Urbana-Champaign (US)
  • National University of Singapore (SG)
  • Center for AI Safety
  • University of Chicago (US)
Venue
Proceedings of the VLDB Endowment
Published
2024-07
Volume
17
Issue
11
Pages
3201-3214
ISSN
2150-8097 (ISSN-L)
OpenAlex
W4402042869
Cited by
42
References
32
Citations by year
2026: 20 · 2025: 21 · 2024: 1
FWCI
6.21
Type
journal article
Language
en
Open access
Locations
  • Proceedings of the VLDB Endowment (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • National Science Foundation (2229876, 2212176)
  • U.S. Department of Homeland Security
  • National Research Foundation
  • National Research Foundation Singapore
  • Info-communications Media Development Authority
  • Defense Advanced Research Projects Agency (2229876)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
  • Physical Sciences › Computer Science › Artificial Intelligence › Adversarial Robustness in Machine Learning
  • Social Sciences › Sociology and Political Science › Privacy, Security, and Data Protection
Keywords
large language models, data privacy, training data leakage, privacy risk assessment, model size

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.