Preprint · 2024 · Open access

LeakAgent: RL-based Red-teaming Agent for LLM Privacy Leakage

Yuzhou Nie, Zhun Wang, Ye Yu, Wu, Xian, Zhao, Xuandong, Wenbo Guo, Dawn Xiaodong Song

arXiv (Cornell University) · Published 7 December 2024 · DOI 10.48550/arxiv.2412.05734

Open PDFView at the publisherCite this paper

In brief

“LeakAgent: RL-based Red-teaming Agent for LLM Privacy Leakage” is a 2024 preprint by Yuzhou Nie and 6 co-authors, published in arXiv (Cornell University). OpenAlex counts 24 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
24
Most cited in
2025 (13)
References
48
Authors
7

Abstract

Recent studies have discovered that large language models (LLM) may be ``fooled'' to output private information, including training data, system prompts, and personally identifiable information, under carefully crafted adversarial prompts. Existing red-teaming approaches for privacy leakage either rely on manual efforts or focus solely on system prompt extraction, making them ineffective for severe risks of training data leakage. We propose LeakAgent, a novel black-box red-teaming framework for LLM privacy leakage. Our framework trains an open-source LLM through reinforcement learning as the attack agent to generate adversarial prompts for both training data extraction and system prompt extraction. To achieve this, we propose a novel reward function to provide effective and fine-grained rewards and design novel mechanisms to balance exploration and exploitation during learning and enhance the diversity of adversarial prompts. Through extensive evaluations, we first show that LeakAgent significantly outperforms existing rule-based approaches in training data extraction and automated methods in system prompt leakage. We also demonstrate the effectiveness of LeakAgent in extracting system prompts from real-world applications in OpenAI's GPT Store. We further demonstrate LeakAgent's effectiveness in evading the existing guardrail defense and its helpfulness in enabling better safety alignment. Finally, we validate our customized designs through a detailed ablation study. We release our code here https://github.com/rucnyz/LeakAgent.

Cite this paper

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

APA 7

Nie, Y., Wang, Z., Yu, Y., Xian, W., Xuandong, Z., Guo, W., & Song, D. X. (2024). LeakAgent: RL-based Red-teaming Agent for LLM Privacy Leakage. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2412.05734

MLA 9

Nie, Yuzhou, et al. “LeakAgent: RL-based Red-teaming Agent for LLM Privacy Leakage.” arXiv (Cornell University), 2024. https://doi.org/10.48550/arxiv.2412.05734.

Chicago

Nie, Yuzhou, Zhun Wang, Ye Yu, Wu, Xian, Zhao, Xuandong, Wenbo Guo, and Dawn Xiaodong Song. 2024. “LeakAgent: RL-based Red-teaming Agent for LLM Privacy Leakage.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2412.05734.

Harvard

Nie, Y. et al. (2024) ‘LeakAgent: RL-based Red-teaming Agent for LLM Privacy Leakage’, arXiv (Cornell University). doi:10.48550/arxiv.2412.05734.

IEEE

Y. Nie et al., “LeakAgent: RL-based Red-teaming Agent for LLM Privacy Leakage,” arXiv (Cornell University), 2024, doi: 10.48550/arxiv.2412.05734.

BibTeX

@misc{nie2024leakagent,
  title     = {{LeakAgent: RL-based Red-teaming Agent for LLM Privacy Leakage}},
  author    = {Nie, Yuzhou and Wang, Zhun and Yu, Ye and Xian, Wu, and Xuandong, Zhao, and Guo, Wenbo and Song, Dawn Xiaodong},
  howpublished = {arXiv (Cornell University)},
  year      = {2024},
  doi       = {10.48550/arxiv.2412.05734},
  url       = {https://doi.org/10.48550/arxiv.2412.05734}
}

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

Citations by year

YearCitationsShare of the best year
202411
20251313
20261010

Topic and keywords

Physical Sciences › Computer Science › Artificial Intelligence › Adversarial Robustness in Machine Learning

  • red teaming
  • training data extraction
  • reinforcement learning agents
  • reward function design
  • exploration-exploitation balance
  • black-box attacks

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?

Nie, Y., Wang, Z., Yu, Y., Xian, W., Xuandong, Z., Guo, W., & Song, D. X. (2024). LeakAgent: RL-based Red-teaming Agent for LLM Privacy Leakage. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2412.05734

How many times has this paper been cited?

24 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?

Yuzhou Nie, Zhun Wang, Ye Yu, Wu, Xian, Zhao, Xuandong, Wenbo Guo, Dawn Xiaodong Song.

What is the DOI of this paper?

10.48550/arxiv.2412.05734, permanently at https://doi.org/10.48550/arxiv.2412.05734. 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.