Private Memorization Editing: Turning Memorization into a Defense to Strengthen Data Privacy in Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ruzzetti, Elena Sofia, Xompero, Giancarlo A., Venditti, Davide, Zanzotto, Fabio Massimo
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915455445237760
author Ruzzetti, Elena Sofia
Xompero, Giancarlo A.
Venditti, Davide
Zanzotto, Fabio Massimo
author_facet Ruzzetti, Elena Sofia
Xompero, Giancarlo A.
Venditti, Davide
Zanzotto, Fabio Massimo
contents Large Language Models (LLMs) memorize, and thus, among huge amounts of uncontrolled data, may memorize Personally Identifiable Information (PII), which should not be stored and, consequently, not leaked. In this paper, we introduce Private Memorization Editing (PME), an approach for preventing private data leakage that turns an apparent limitation, that is, the LLMs' memorization ability, into a powerful privacy defense strategy. While attacks against LLMs have been performed exploiting previous knowledge regarding their training data, our approach aims to exploit the same kind of knowledge in order to make a model more robust. We detect a memorized PII and then mitigate the memorization of PII by editing a model knowledge of its training data. We verify that our procedure does not affect the underlying language model while making it more robust against privacy Training Data Extraction attacks. We demonstrate that PME can effectively reduce the number of leaked PII in a number of configurations, in some cases even reducing the accuracy of the privacy attacks to zero.
format Preprint
id arxiv_https___arxiv_org_abs_2506_10024
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Private Memorization Editing: Turning Memorization into a Defense to Strengthen Data Privacy in Large Language Models
Ruzzetti, Elena Sofia
Xompero, Giancarlo A.
Venditti, Davide
Zanzotto, Fabio Massimo
Cryptography and Security
Artificial Intelligence
Computation and Language
Large Language Models (LLMs) memorize, and thus, among huge amounts of uncontrolled data, may memorize Personally Identifiable Information (PII), which should not be stored and, consequently, not leaked. In this paper, we introduce Private Memorization Editing (PME), an approach for preventing private data leakage that turns an apparent limitation, that is, the LLMs' memorization ability, into a powerful privacy defense strategy. While attacks against LLMs have been performed exploiting previous knowledge regarding their training data, our approach aims to exploit the same kind of knowledge in order to make a model more robust. We detect a memorized PII and then mitigate the memorization of PII by editing a model knowledge of its training data. We verify that our procedure does not affect the underlying language model while making it more robust against privacy Training Data Extraction attacks. We demonstrate that PME can effectively reduce the number of leaked PII in a number of configurations, in some cases even reducing the accuracy of the privacy attacks to zero.
title Private Memorization Editing: Turning Memorization into a Defense to Strengthen Data Privacy in Large Language Models
topic Cryptography and Security
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2506.10024