Towards Evaluation for Real-World LLM Unlearning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Miao, Ke, Hu, Yuke, Li, Xiaochen, Bao, Wenjie, Liu, Zhihao, Qin, Zhan, Ren, Kui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916877709606912
author Miao, Ke
Hu, Yuke
Li, Xiaochen
Bao, Wenjie
Liu, Zhihao
Qin, Zhan
Ren, Kui
author_facet Miao, Ke
Hu, Yuke
Li, Xiaochen
Bao, Wenjie
Liu, Zhihao
Qin, Zhan
Ren, Kui
contents This paper analyzes the limitations of existing unlearning evaluation metrics in terms of practicality, exactness, and robustness in real-world LLM unlearning scenarios. To overcome these limitations, we propose a new metric called Distribution Correction-based Unlearning Evaluation (DCUE). It identifies core tokens and corrects distributional biases in their confidence scores using a validation set. The evaluation results are quantified using the Kolmogorov-Smirnov test. Experimental results demonstrate that DCUE overcomes the limitations of existing metrics, which also guides the design of more practical and reliable unlearning algorithms in the future.
format Preprint
id arxiv_https___arxiv_org_abs_2508_01324
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards Evaluation for Real-World LLM Unlearning
Miao, Ke
Hu, Yuke
Li, Xiaochen
Bao, Wenjie
Liu, Zhihao
Qin, Zhan
Ren, Kui
Artificial Intelligence
This paper analyzes the limitations of existing unlearning evaluation metrics in terms of practicality, exactness, and robustness in real-world LLM unlearning scenarios. To overcome these limitations, we propose a new metric called Distribution Correction-based Unlearning Evaluation (DCUE). It identifies core tokens and corrects distributional biases in their confidence scores using a validation set. The evaluation results are quantified using the Kolmogorov-Smirnov test. Experimental results demonstrate that DCUE overcomes the limitations of existing metrics, which also guides the design of more practical and reliable unlearning algorithms in the future.
title Towards Evaluation for Real-World LLM Unlearning
topic Artificial Intelligence
url https://arxiv.org/abs/2508.01324