Cross-Modal VLM Unlearning Vulnerabilities

Arxiv pdf 2026-07-01T00:00:00
arXiv Paper — PDF not available. Only the Executive Summary is available here. To read or download the full paper, visit the arXiv abstract page.

Abstract

Modern Vision-Language Models (VLMs), however, process both text and visual inputs, raising a fundamental security question: does unlearning in one modality transfer to the other? We present the first systematic, bidirectional study of cross-modal unlearning transfer across three VLM architectures: LLaVA-1.5 (MLP projection), InstructBLIP (Q-Former), and IDEFICS (gated cross-attention). We find that unlearning transfers across modalities, but the transfer is asymmetric and incomplete. In some cases, text unlearning strongly transfers to vision. However, this robustness is not preserved under typographic attacks that manipulate the visual presentation of text. Under such attacks, previously unlearned knowledge can be readily recovered, indicating shallow unlearning.

Loading executive summary...

LINK COPIED TO CLIPBOARD