EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Differential Privacy-Based Mechanism for Preventing Data Leakage in Large Language Model Training

Xingpeng Xiao, Yaomin Zhang, Heyao Chen, Wenkun Ren, Junyi Zhang, Jian Xu

Academic Journal of Sociology and Management · 2025

DOI: 10.70393/616a736d.323732 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Academic Journal of Sociology and Management, 2025. Verfasst von Xingpeng Xiao und 5 weiteren Autor:innen. Die Arbeit wurde bisher 7-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large Language Models (LLMs) have demonstrated remarkable capabilities in natural language processing tasks, yet they face significant challenges in protecting sensitive information during training. This paper presents a novel differential privacy-based mechanism for preventing data leakage in LLM training processes. The proposed system introduces a dynamic privacy budget allocation strategy integrated with adaptive noise injection mechanisms, specifically designed for transformer architectures. The mechanism implements a multi-layered protection framework that combines real-time monitoring capabilities with automated response systems. Through comprehensive experimental evaluation on models ranging from 100M to 175B parameters, our approach demonstrates superior performance in privacy protection while maintaining model utility. The system achieves a 99.2% detection rate for potential data leakages with a minimal false alarm rate of 0.8%, representing a significant improvement over traditional approaches. Performance analysis reveals that the proposed mechanism maintains model accuracy within 1.8% of non-private baselines while providing strong privacy guarantees. The implementation reduces computational overhead by 35% compared to conventional differential privacy methods. Our research establishes new benchmarks in privacy-preserving machine learning, particularly for large-scale language models, and provides a practical framework for secure AI system deployment.

Zitieren

APA

Xiao, X., Zhang, Y., Chen, H., Ren, W., Zhang, J., & Xu, J. (2025). A Differential Privacy-Based Mechanism for Preventing Data Leakage in Large Language Model Training. Academic Journal of Sociology and Management. https://doi.org/10.70393/616a736d.323732

Vancouver

Xiao X, Zhang Y, Chen H, Ren W, Zhang J, Xu J. A Differential Privacy-Based Mechanism for Preventing Data Leakage in Large Language Model Training. Academic Journal of Sociology and Management. 2025. doi:10.70393/616a736d.323732

BibTeX
@article{Xiao2025,
  author = {Xiao, Xingpeng and Zhang, Yaomin and Chen, Heyao and Ren, Wenkun and Zhang, Junyi and Xu, Jian},
  title = {A Differential Privacy-Based Mechanism for Preventing Data Leakage in Large Language Model Training},
  journal = {Academic Journal of Sociology and Management},
  year = {2025},
  doi = {10.70393/616a736d.323732}
}
Kostenlos starten →