Preprint · 2023 · Open access
Harnessing large-language models to generate private synthetic text
arXiv (Cornell University) · Published 2 June 2023 · DOI 10.48550/arxiv.2306.01684
In brief
“Harnessing large-language models to generate private synthetic text” is a 2023 preprint by Alexey Kurakin and 4 co-authors, published in arXiv (Cornell University). OpenAlex counts 103 citations as of 19 November 2026. It is open access: the full text is free to read at arXiv (Cornell University).
- Citations
- 103
- Most cited in
- 2025 (46)
- References
- 43
- Authors
- 5
Abstract
Differentially private training algorithms like DP-SGD protect sensitive training data by ensuring that trained models do not reveal private information. An alternative approach, which this paper studies, is to use a sensitive dataset to generate synthetic data that is differentially private with respect to the original data, and then non-privately training a model on the synthetic data. Doing so has several advantages: synthetic data can be reused for other tasks (including for hyper parameter tuning), retained indefinitely, and shared with third parties without sacrificing privacy. However, generating private synthetic data is much harder than training a private model. To improve performance on text data, recent work has utilized public data by starting with a pre-trained generative language model and privately fine-tuning it on sensitive data. This model can be used to sample a DP synthetic dataset. While this strategy seems straightforward, executing it has proven problematic. Previous approaches either show significant performance loss, or have, as we show, critical design flaws. In this paper we demonstrate that a proper training objective along with tuning fewer parameters results in excellent DP synthetic data quality. Our approach is competitive with direct DP-training of downstream classifiers in terms of performance on downstream tasks. Further, we demonstrate that our DP synthetic data is not only useful for downstream classifier training, but also to tune those same models.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Kurakin, A., Ponomareva, N., Umar, S., MacDermed, L., & Terzis, A. (2023). Harnessing large-language models to generate private synthetic text. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2306.01684
MLA 9
Kurakin, Alexey, et al. “Harnessing large-language models to generate private synthetic text.” arXiv (Cornell University), 2023. https://doi.org/10.48550/arxiv.2306.01684.
Chicago
Kurakin, Alexey, Natalia Ponomareva, Syed, Umar, Liam MacDermed, and Andreas Terzis. 2023. “Harnessing large-language models to generate private synthetic text.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2306.01684.
Harvard
Kurakin, A. et al. (2023) ‘Harnessing large-language models to generate private synthetic text’, arXiv (Cornell University). doi:10.48550/arxiv.2306.01684.
IEEE
A. Kurakin, N. Ponomareva, S. Umar, L. MacDermed, and A. Terzis, “Harnessing large-language models to generate private synthetic text,” arXiv (Cornell University), 2023, doi: 10.48550/arxiv.2306.01684.
BibTeX
@misc{kurakin2023harnessing,
title = {{Harnessing large-language models to generate private synthetic text}},
author = {Kurakin, Alexey and Ponomareva, Natalia and Umar, Syed, and MacDermed, Liam and Terzis, Andreas},
howpublished = {arXiv (Cornell University)},
year = {2023},
doi = {10.48550/arxiv.2306.01684},
url = {https://doi.org/10.48550/arxiv.2306.01684}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Citations by year
| Year | Citations | Share of the best year |
|---|---|---|
| 2021 | 1 | |
| 2022 | 1 | |
| 2023 | 3 | |
| 2024 | 19 | |
| 2025 | 46 | |
| 2026 | 33 |
Topic and keywords
Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- differential privacy
- synthetic data generation
- large language models
- DP-SGD
- fine-tuning
- hyperparameter optimization
- data privacy
Work with this paper
Most-cited papers in Computer Science
- ImageNet classification with deep convolutional neural networksAlex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton · Communications of the ACM · 2017 · 114,726 citations
- Long Short-Term MemorySepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 109,097 citations
- A Mathematical Theory of CommunicationClaude E. Shannon · Bell System Technical Journal · 1948 · 91,736 citations
- Deep learningYann LeCun, Yoshua Bengio, Geoffrey E. Hinton · Nature · 2015 · 90,226 citations
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksShaoqing Ren, Kaiming He, Ross Girshick and 1 more · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2016 · 74,143 citations
Questions about this paper
How do I cite this paper in APA style?
Kurakin, A., Ponomareva, N., Umar, S., MacDermed, L., & Terzis, A. (2023). Harnessing large-language models to generate private synthetic text. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2306.01684
How many times has this paper been cited?
103 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access at arXiv (Cornell University): sign in and open the PDF from this page.
Who wrote this paper?
Alexey Kurakin, Natalia Ponomareva, Syed, Umar, Liam MacDermed, Andreas Terzis.
What is the DOI of this paper?
10.48550/arxiv.2306.01684, permanently at https://doi.org/10.48550/arxiv.2306.01684. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.