Papers
arxiv:2506.01339

Invariance Makes LLM Unlearning Resilient Even to Unanticipated Downstream Fine-Tuning

Published on Oct 10, 2025
Authors:
,
,
,
,
,
,
,
,

Abstract

Invariant machine unlearning framework enhances robustness against knowledge recovery in large language models through regularization and generalizes across diverse downstream tasks.

AI-generated summary

Machine unlearning offers a promising solution to privacy and safety concerns in large language models (LLMs) by selectively removing targeted knowledge while preserving utility. However, current methods are highly sensitive to downstream fine-tuning, which can quickly recover forgotten information-even from unrelated tasks. To address this, we introduce invariance into unlearning for the first time, inspired by invariant risk minimization (IRM). Building on this principle, we propose invariant LLM unlearning (ILU), a regularization-based framework that enhances robustness. Notably, ILU generalizes well to diverse fine-tuning tasks, even when trained using a single dataset. A task vector analysis is also provided to further elucidate the rationale behind ILU's effectiveness. Extensive experiments on the WMDP and MUSE benchmark, reveal that ILU significantly outperforms state-of-the-art unlearning methods, including negative preference optimization (NPO) and representation misdirection for unlearning (RMU). Notably, ILU achieves superior unlearning robustness across diverse downstream fine-tuning scenarios (e.g., math, paraphrase detection, and sentiment analysis) while preserving the fine-tuning performance.

Community

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2506.01339
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2506.01339 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2506.01339 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2506.01339 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.