arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.20580cs.CLcs.AIcs.CRcs.LG
收藏

语言模型中的个人信息重复

Personal Information Parroting in Language Models

Carnegie Mellon University - Language Technologies Institute(卡内基梅隆大学-语言技术研究所);University of Washington(华盛顿大学)

Nishant Subramani, Kshitish Ghate, Mona Diab

AI总结:

本研究开发了检测个人信息的正则表达式和规则套件,发现语言模型会记忆并重复个人信息,建议加强数据过滤和匿名化以降低隐私风险。

摘要:

现代语言模型(LM)是在大量的网络数据上进行训练的,其中包含数以百万计的个人信息(PI)实例,许多实例被LM记忆化,增加了隐私风险。在本工作中,我们开发了正则表达式和规则(R&R)检测器套件,用于检测电子邮件地址、电话号码和IP地址,其性能优于最佳的基于正则表达式的PI检测器。在手动整理的483个PI实例集上,我们测量了记忆化现象:发现Pythia-6.9b模型中13.6%的实例被逐字重复,即当模型被提示以原文档中PI前的token时,贪心解码会生成完整的PI跨度。我们将此分析扩展到研究Pythia模型套件中不同大小(160M-6.9B)和预训练时间步(70k-143k迭代)的模型,并发现模型大小和预训练量与记忆化呈正相关。即使最小的模型Pythia-160m也精确重复了2.7%的实例。因此,我们强烈建议预训练数据集应被积极过滤和匿名化,以最小化PI重复。

英文摘要:

Modern language models (LM) are trained on large scrapes of the Web, containing millions of personal information (PI) instances, many of which LMs memorize, increasing privacy risks. In this work, we develop the regexes and rules (R&R) detector suite to detect email addresses, phone numbers, and IP addresses, which outperforms the best regex-based PI detectors. On a manually curated set of 483 instances of PI, we measure memorization: finding that 13.6% are parroted verbatim by the Pythia-6.9b model, i.e., when the model is prompted with the tokens that precede the PI in the original document, greedy decoding generates the entire PI span exactly. We expand this analysis to study models of varying sizes (160M-6.9B) and pretraining time steps (70k-143k iterations) in the Pythia model suite and find that both model size and amount of pretraining are positively correlated with memorization. Even the smallest model, Pythia-160m, parrots 2.7% of the instances exactly. Consequently, we strongly recommend that pretraining datasets be aggressively filtered and anonymized to minimize PI parroting.

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。

↑