Locally Differentially Private Document Generation Using Zero Shot Prompting

2023/10/24 by Saiteja Utpala, Sara Hooker, Utpala, Saiteja +3 · 11 citations
Computer Science · Medicine · #Artificial Intelligence in Healthcare and Education #Computation and Language (cs.CL) #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Privacy-Preserving Technologies in Data #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2310.16111

openalex publication_date 2023/10/24 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Numerous studies have highlighted the privacy risks associated with pretrained large language models. In contrast, our research offers a unique perspective by demonstrating that pretrained large language models can effectively contribute to privacy preservation. We propose a locally differentially private mechanism called DP-Prompt, which leverages the power of pretrained large language models and zero-shot prompting to counter author de-anonymization attacks while minimizing the impact on downstream utility. When DP-Prompt is used with a powerful language model like ChatGPT (gpt-3.5), we observe a notable reduction in the success rate of de-anonymization attacks, showing that it surpasses existing approaches by a considerable margin despite its simpler design. For instance, in the case of the IMDB dataset, DP-Prompt (with ChatGPT) perfectly recovers the clean sentiment F1 score while achieving a 46% reduction in author identification F1 score against static attackers and a 26% reduction against adaptive attackers. We conduct extensive experiments across six open-source large language models, ranging up to 7 billion parameters, to analyze various effects of the privacy-utility tradeoff.

Cited by

• PrivCode: When Code Generation Meets Differential Privacy
• With Privacy, Size Matters: On the Importance of Dataset Size in Differentially Private Text Rewriting
• One Size Fits All? A Modular Adaptive Sanitization Kit (MASK) for Customizable Privacy-Preserving Phone Scam Detection
• Controlled Generation for Private Synthetic Text
• Zero-Shot Privacy-Aware Text Rewriting via Iterative Tree Search
• RL-Finetuned LLMs for Privacy-Preserving Synthetic Rewriting
• Privacy Preserving In-Context-Learning Framework for Large Language Models
• Leveraging Semantic Triples for Private Document Generation with Local Differential Privacy Guarantees
• The Double-edged Sword of LLM-based Data Reconstruction: Understanding and Mitigating Contextual Vulnerability in Word-level Differential Privacy Text Sanitization
• LLM-as-a-Judge for Privacy Evaluation? Exploring the Alignment of Human and LLM Perceptions of Privacy in Textual Data
• When Explainability Meets Privacy: An Investigation at the Intersection of Post-hoc Explainability and Differential Privacy in the Context of Natural Language Processing

Related

• Reconstruction of Differentially Private Text Sanitization via Large Language Models
• InferDPT: Privacy-Preserving Inference for Black-box Large Language Model
• You Are What You Write: Preserving Privacy in the Era of Large Language Models
• DP-Fusion: Token-Level Differentially Private Inference for Large Language Models
• DP-GTR: Differentially Private Prompt Protection via Group Text Rewriting