Enhancing Data Privacy in Large Language Models through Private Association Editing
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910652540387328 |
|---|---|
| author | Venditti, Davide Ruzzetti, Elena Sofia Xompero, Giancarlo A. Giannone, Cristina Favalli, Andrea Romagnoli, Raniero Zanzotto, Fabio Massimo |
| author_facet | Venditti, Davide Ruzzetti, Elena Sofia Xompero, Giancarlo A. Giannone, Cristina Favalli, Andrea Romagnoli, Raniero Zanzotto, Fabio Massimo |
| contents | Large language models (LLMs) require a significant redesign in solutions to preserve privacy in data-intensive applications due to their text-generation capabilities. Indeed, LLMs tend to memorize and emit private information when maliciously prompted. In this paper, we introduce Private Association Editing (PAE) as a novel defense approach for private data leakage. PAE is designed to effectively remove Personally Identifiable Information (PII) without retraining the model. Experimental results demonstrate the effectiveness of PAE with respect to alternative baseline methods. We believe PAE will serve as a critical tool in the ongoing effort to protect data privacy in LLMs, encouraging the development of safer models for real-world applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_18221 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Enhancing Data Privacy in Large Language Models through Private Association Editing Venditti, Davide Ruzzetti, Elena Sofia Xompero, Giancarlo A. Giannone, Cristina Favalli, Andrea Romagnoli, Raniero Zanzotto, Fabio Massimo Computation and Language Artificial Intelligence Large language models (LLMs) require a significant redesign in solutions to preserve privacy in data-intensive applications due to their text-generation capabilities. Indeed, LLMs tend to memorize and emit private information when maliciously prompted. In this paper, we introduce Private Association Editing (PAE) as a novel defense approach for private data leakage. PAE is designed to effectively remove Personally Identifiable Information (PII) without retraining the model. Experimental results demonstrate the effectiveness of PAE with respect to alternative baseline methods. We believe PAE will serve as a critical tool in the ongoing effort to protect data privacy in LLMs, encouraging the development of safer models for real-world applications. |
| title | Enhancing Data Privacy in Large Language Models through Private Association Editing |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2406.18221 |