Preprint · 2024 · Open access

LMO-DP: Optimizing the Randomization Mechanism for Differentially Private Fine-Tuning (Large) Language Models

Qin Yang, Meisam Mohammad, Han Wang, Ali Payani, Ashish Kundu, Kai Shu, Yan Yan, Yuan Hong

arXiv (Cornell University) · Published 29 May 2024 · DOI 10.48550/arxiv.2405.18776

Open PDFView at the publisherCite this paper

In brief

“LMO-DP: Optimizing the Randomization Mechanism for Differentially Private Fine-Tuning (Large) Language Models” is a 2024 preprint by Qin Yang and 7 co-authors, published in arXiv (Cornell University). OpenAlex counts 3 citations as of 19 November 2026. It is open access: the full text is free to read at arXiv (Cornell University).

Citations
3
Most cited in
2025 (2)
References
20
Authors
8

Abstract

Differentially Private Stochastic Gradient Descent (DP-SGD) and its variants have been proposed to ensure rigorous privacy for fine-tuning large-scale pre-trained language models. However, they rely heavily on the Gaussian mechanism, which may overly perturb the gradients and degrade the accuracy, especially in stronger privacy regimes (e.g., the privacy budget $ε< 3$). To address such limitations, we propose a novel Language Model-based Optimal Differential Privacy (LMO-DP) mechanism, which takes the first step to enable the tight composition of accurately fine-tuning (large) language models with a sub-optimal DP mechanism, even in strong privacy regimes (e.g., $0.1\leq ε<3$). Furthermore, we propose a novel offline optimal noise search method to efficiently derive the sub-optimal DP that significantly reduces the noise magnitude. For instance, fine-tuning RoBERTa-large (with 300M parameters) on the SST-2 dataset can achieve an accuracy of 92.20% (given $ε=0.3$, $δ=10^{-10}$) by drastically outperforming the Gaussian mechanism (e.g., $\sim 50\%$ for small $ε$ and $δ$). We also draw similar findings on the text generation tasks on GPT-2. Finally, to our best knowledge, LMO-DP is also the first solution to accurately fine-tune Llama-2 with strong differential privacy guarantees. The code will be released soon and available upon request.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Yang, Q., Mohammad, M., Wang, H., Payani, A., Kundu, A., Shu, K., Yan, Y., & Hong, Y. (2024). LMO-DP: Optimizing the Randomization Mechanism for Differentially Private Fine-Tuning (Large) Language Models. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2405.18776

MLA 9

Yang, Qin, et al. “LMO-DP: Optimizing the Randomization Mechanism for Differentially Private Fine-Tuning (Large) Language Models.” arXiv (Cornell University), 2024. https://doi.org/10.48550/arxiv.2405.18776.

Chicago

Yang, Qin, Meisam Mohammad, Han Wang, Ali Payani, Ashish Kundu, Kai Shu, Yan Yan, and Yuan Hong. 2024. “LMO-DP: Optimizing the Randomization Mechanism for Differentially Private Fine-Tuning (Large) Language Models.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2405.18776.

Harvard

Yang, Q. et al. (2024) ‘LMO-DP: Optimizing the Randomization Mechanism for Differentially Private Fine-Tuning (Large) Language Models’, arXiv (Cornell University). doi:10.48550/arxiv.2405.18776.

IEEE

Q. Yang et al., “LMO-DP: Optimizing the Randomization Mechanism for Differentially Private Fine-Tuning (Large) Language Models,” arXiv (Cornell University), 2024, doi: 10.48550/arxiv.2405.18776.

BibTeX

@misc{yang2024optimizing,
  title     = {{LMO-DP: Optimizing the Randomization Mechanism for Differentially Private Fine-Tuning (Large) Language Models}},
  author    = {Yang, Qin and Mohammad, Meisam and Wang, Han and Payani, Ali and Kundu, Ashish and Shu, Kai and Yan, Yan and Hong, Yuan},
  howpublished = {arXiv (Cornell University)},
  year      = {2024},
  doi       = {10.48550/arxiv.2405.18776},
  url       = {https://doi.org/10.48550/arxiv.2405.18776}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Topic and keywords

Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data

  • differential privacy
  • language model fine-tuning
  • DP-SGD
  • Gaussian mechanism
  • privacy-utility trade-off
  • large language models
  • text generation
  • Llama 2

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Yang, Q., Mohammad, M., Wang, H., Payani, A., Kundu, A., Shu, K., Yan, Y., & Hong, Y. (2024). LMO-DP: Optimizing the Randomization Mechanism for Differentially Private Fine-Tuning (Large) Language Models. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2405.18776

How many times has this paper been cited?

3 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?

Qin Yang, Meisam Mohammad, Han Wang, Ali Payani, Ashish Kundu, Kai Shu, Yan Yan, Yuan Hong.

What is the DOI of this paper?

10.48550/arxiv.2405.18776, permanently at https://doi.org/10.48550/arxiv.2405.18776. 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.