Conference paper · 2025 · Open access

Can We Infer Confidential Properties of Training Data from LLMs?

Pengrun Huang, Chhavi Yadav, Kamalika Chaudhuri, Ruihan Wu

neural information processing systems, 70537–70561 · Published 1 January 2025 · DOI 10.52202/085713-2124

Open PDFView at the publisherCite this paper

In brief

“Can We Infer Confidential Properties of Training Data from LLMs?” is a 2025 conference paper by Pengrun Huang and 3 co-authors, published in neural information processing systems. OpenAlex counts 8 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at arXiv (Cornell University).

Citations
8
Most cited in
2026 (7)
Field-weighted impact
1.66
References
20
Authors
4

Abstract

Large language models (LLMs) are increasingly fine-tuned on domain-specific datasets to support applications in fields such as healthcare, finance, and law. These fine-tuning datasets often have sensitive and confidential dataset-level properties -- such as patient demographics or disease prevalence -- that are not intended to be revealed. While prior work has studied property inference attacks on discriminative models (e.g., image classification models) and generative models (e.g., GANs for image data), it remains unclear if such attacks transfer to LLMs. In this work, we introduce PropInfer, a benchmark task for evaluating property inference in LLMs under two fine-tuning paradigms: question-answering and chat-completion. Built on the ChatDoctor dataset, our benchmark includes a range of property types and task configurations. We further propose two tailored attacks: a prompt-based generation attack and a shadow-model attack leveraging word frequency signals. Empirical evaluations across multiple pretrained LLMs show the success of our attacks, revealing a previously unrecognized vulnerability in LLMs.

Cite this paper

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

APA 7

Huang, P., Yadav, C., Chaudhuri, K., & Wu, R. (2025). Can We Infer Confidential Properties of Training Data from LLMs? neural information processing systems, 70537–70561. https://doi.org/10.52202/085713-2124

MLA 9

Huang, Pengrun, et al. “Can We Infer Confidential Properties of Training Data from LLMs?” neural information processing systems, 2025, pp. 70537–70561. https://doi.org/10.52202/085713-2124.

Chicago

Huang, Pengrun, Chhavi Yadav, Kamalika Chaudhuri, and Ruihan Wu. 2025. “Can We Infer Confidential Properties of Training Data from LLMs?” neural information processing systems: 70537–70561. https://doi.org/10.52202/085713-2124.

Harvard

Huang, P. et al. (2025) ‘Can We Infer Confidential Properties of Training Data from LLMs?’, neural information processing systems, pp. 70537–70561. doi:10.52202/085713-2124.

IEEE

P. Huang, C. Yadav, K. Chaudhuri, and R. Wu, “Can We Infer Confidential Properties of Training Data from LLMs?,” neural information processing systems, pp. 70537–70561, 2025, doi: 10.52202/085713-2124.

BibTeX

@inproceedings{huang2025infer,
  title     = {{Can We Infer Confidential Properties of Training Data from LLMs?}},
  author    = {Huang, Pengrun and Yadav, Chhavi and Chaudhuri, Kamalika and Wu, Ruihan},
  booktitle = {neural information processing systems},
  year      = {2025},
  pages     = {70537--70561},
  doi       = {10.52202/085713-2124},
  url       = {https://doi.org/10.52202/085713-2124}
}

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

  • property inference attacks
  • large language models
  • data privacy
  • fine-tuning
  • benchmark
  • question answering

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?

Huang, P., Yadav, C., Chaudhuri, K., & Wu, R. (2025). Can We Infer Confidential Properties of Training Data from LLMs? neural information processing systems, 70537–70561. https://doi.org/10.52202/085713-2124

How many times has this paper been cited?

8 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?

Pengrun Huang, Chhavi Yadav, Kamalika Chaudhuri, Ruihan Wu.

What is the DOI of this paper?

10.52202/085713-2124, permanently at https://doi.org/10.52202/085713-2124. 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.