← Search

Nael B. Abu-Ghazaleh

2 accepted papers

2025

Layer-wise Alignment: Examining Safety Alignment Across Image Encoder Layers in Vision Language Models

ICML 2025spotlight

Vision-language models (VLMs) have improved significantly in their capabilities, but their complex architecture makes their safety alignment challenging. In this paper, we reveal an uneven distribution of harmful information across the intermediate layers of the image encoder and show that skipping…

Cited by 0SourcePDFScholar
2024

Can Textual Unlearning Solve Cross-Modality Safety Alignment?

EMNLP 2024finding

Recent studies reveal that integrating new modalities into large language models (LLMs), such as vision-language models (VLMs), creates a new attack surface that bypasses existing safety training techniques like supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF). Whil…

Cited by 1SourcePDFScholar