Unlocking Memorization in Large Language Models with Dynamic Soft Prompting
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913511882358784 |
|---|---|
| author | Wang, Zhepeng Bao, Runxue Wu, Yawen Taylor, Jackson Xiao, Cao Zheng, Feng Jiang, Weiwen Gao, Shangqian Zhang, Yanfu |
| author_facet | Wang, Zhepeng Bao, Runxue Wu, Yawen Taylor, Jackson Xiao, Cao Zheng, Feng Jiang, Weiwen Gao, Shangqian Zhang, Yanfu |
| contents | Pretrained large language models (LLMs) have revolutionized natural language processing (NLP) tasks such as summarization, question answering, and translation. However, LLMs pose significant security risks due to their tendency to memorize training data, leading to potential privacy breaches and copyright infringement. Accurate measurement of this memorization is essential to evaluate and mitigate these potential risks. However, previous attempts to characterize memorization are constrained by either using prefixes only or by prepending a constant soft prompt to the prefixes, which cannot react to changes in input. To address this challenge, we propose a novel method for estimating LLM memorization using dynamic, prefix-dependent soft prompts. Our approach involves training a transformer-based generator to produce soft prompts that adapt to changes in input, thereby enabling more accurate extraction of memorized data. Our method not only addresses the limitations of previous methods but also demonstrates superior performance in diverse experimental settings compared to state-of-the-art techniques. In particular, our method can achieve the maximum relative improvement of 112.75% and 32.26% over the vanilla baseline in terms of discoverable memorization rate for the text generation task and code generation task respectively. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_13853 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Unlocking Memorization in Large Language Models with Dynamic Soft Prompting Wang, Zhepeng Bao, Runxue Wu, Yawen Taylor, Jackson Xiao, Cao Zheng, Feng Jiang, Weiwen Gao, Shangqian Zhang, Yanfu Computation and Language Artificial Intelligence Cryptography and Security Machine Learning Pretrained large language models (LLMs) have revolutionized natural language processing (NLP) tasks such as summarization, question answering, and translation. However, LLMs pose significant security risks due to their tendency to memorize training data, leading to potential privacy breaches and copyright infringement. Accurate measurement of this memorization is essential to evaluate and mitigate these potential risks. However, previous attempts to characterize memorization are constrained by either using prefixes only or by prepending a constant soft prompt to the prefixes, which cannot react to changes in input. To address this challenge, we propose a novel method for estimating LLM memorization using dynamic, prefix-dependent soft prompts. Our approach involves training a transformer-based generator to produce soft prompts that adapt to changes in input, thereby enabling more accurate extraction of memorized data. Our method not only addresses the limitations of previous methods but also demonstrates superior performance in diverse experimental settings compared to state-of-the-art techniques. In particular, our method can achieve the maximum relative improvement of 112.75% and 32.26% over the vanilla baseline in terms of discoverable memorization rate for the text generation task and code generation task respectively. |
| title | Unlocking Memorization in Large Language Models with Dynamic Soft Prompting |
| topic | Computation and Language Artificial Intelligence Cryptography and Security Machine Learning |
| url | https://arxiv.org/abs/2409.13853 |