arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.09266cs.CRcs.CLcs.LG
收藏

针对大语言模型的用户推断攻击

User Inference Attacks on Large Language Models

University of Toronto(多伦多大学);Vector Institute(矢量研究院);Google(谷歌公司);Northeastern University(东北大学)

Nikhil Kandpal, Krishna Pillutla, Alina Oprea, Peter Kairouz, Christopher A. Choquette-Choo, Zheng Xu

AI总结:

本文研究大语言模型微调中的用户推断隐私攻击,提出仅需黑盒访问和少量用户样本的攻击方法,发现离群用户等更易受攻击,并探讨了差分隐私等缓解措施。

摘要:

微调是一种常见且有效的方法,用于将大语言模型(LLMs)适配到专门的任务和应用中。本文研究在用户数据上微调LLMs的隐私影响。为此,我们考虑一种现实的威胁模型,称为用户推断,其中攻击者推断某个用户的数据是否被用于微调。我们设计了执行用户推断的攻击方法,这些方法仅需要对微调后的LLM进行黑盒访问,以及来自某个用户的少量样本,且这些样本不必来自微调数据集。我们发现,LLMs在多种微调数据集上都容易受到用户推断攻击,有时攻击成功率接近完美。此外,我们从理论和实证上研究了使用户易受用户推断攻击的特性,发现离群用户、样本之间具有可识别共享特征的用户,以及贡献了微调数据中很大一部分的用户最容易受到攻击。基于这些发现,我们确定了几种缓解用户推断的方法,包括使用样本级差分隐私进行训练、移除用户内部的重复样本,以及减少用户对训练数据的贡献。虽然这些技术能够部分缓解用户推断,但我们强调需要开发能够完全保护微调后LLMs免受这一隐私风险的方法。

英文摘要:

Fine-tuning is a common and effective method for tailoring large language models (LLMs) to specialized tasks and applications. In this paper, we study the privacy implications of fine-tuning LLMs on user data. To this end, we consider a realistic threat model, called user inference, wherein an attacker infers whether or not a user's data was used for fine-tuning. We design attacks for performing user inference that require only black-box access to the fine-tuned LLM and a few samples from a user which need not be from the fine-tuning dataset. We find that LLMs are susceptible to user inference across a variety of fine-tuning datasets, at times with near perfect attack success rates. Further, we theoretically and empirically investigate the properties that make users vulnerable to user inference, finding that outlier users, users with identifiable shared features between examples, and users that contribute a large fraction of the fine-tuning data are most susceptible to attack. Based on these findings, we identify several methods for mitigating user inference including training with example-level differential privacy, removing within-user duplicate examples, and reducing a user's contribution to the training data. While these techniques provide partial mitigation of user inference, we highlight the need to develop methods to fully protect fine-tuned LLMs against this privacy risk.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑