Probing Language Models for Pre-training Data Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Zhenhua, Zhu, Tong, Tan, Chuanyuan, Lu, Haonan, Liu, Bing, Chen, Wenliang
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914820961337344
author Liu, Zhenhua
Zhu, Tong
Tan, Chuanyuan
Lu, Haonan
Liu, Bing
Chen, Wenliang
author_facet Liu, Zhenhua
Zhu, Tong
Tan, Chuanyuan
Lu, Haonan
Liu, Bing
Chen, Wenliang
contents Large Language Models (LLMs) have shown their impressive capabilities, while also raising concerns about the data contamination problems due to privacy issues and leakage of benchmark datasets in the pre-training phase. Therefore, it is vital to detect the contamination by checking whether an LLM has been pre-trained on the target texts. Recent studies focus on the generated texts and compute perplexities, which are superficial features and not reliable. In this study, we propose to utilize the probing technique for pre-training data detection by examining the model's internal activations. Our method is simple and effective and leads to more trustworthy pre-training data detection. Additionally, we propose ArxivMIA, a new challenging benchmark comprising arxiv abstracts from Computer Science and Mathematics categories. Our experiments demonstrate that our method outperforms all baselines, and achieves state-of-the-art performance on both WikiMIA and ArxivMIA, with additional experiments confirming its efficacy (Our code and dataset are available at https://github.com/zhliu0106/probing-lm-data).
format Preprint
id arxiv_https___arxiv_org_abs_2406_01333
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Probing Language Models for Pre-training Data Detection
Liu, Zhenhua
Zhu, Tong
Tan, Chuanyuan
Lu, Haonan
Liu, Bing
Chen, Wenliang
Computation and Language
Artificial Intelligence
Large Language Models (LLMs) have shown their impressive capabilities, while also raising concerns about the data contamination problems due to privacy issues and leakage of benchmark datasets in the pre-training phase. Therefore, it is vital to detect the contamination by checking whether an LLM has been pre-trained on the target texts. Recent studies focus on the generated texts and compute perplexities, which are superficial features and not reliable. In this study, we propose to utilize the probing technique for pre-training data detection by examining the model's internal activations. Our method is simple and effective and leads to more trustworthy pre-training data detection. Additionally, we propose ArxivMIA, a new challenging benchmark comprising arxiv abstracts from Computer Science and Mathematics categories. Our experiments demonstrate that our method outperforms all baselines, and achieves state-of-the-art performance on both WikiMIA and ArxivMIA, with additional experiments confirming its efficacy (Our code and dataset are available at https://github.com/zhliu0106/probing-lm-data).
title Probing Language Models for Pre-training Data Detection
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2406.01333