VLM safety training flawed by spurious correlations, study finds

Researchers have identified a significant flaw in current safety training for vision-language models (VLMs), termed the "safety mirage." This occurs when models learn spurious correlations between superficial text patterns and safety responses, rather than truly understanding harm. These VLMs can be easily tricked by simple word substitutions, leading to bypassed safeguards or unnecessary rejections of benign queries. The study proposes machine unlearning (MU) as a more effective method for safety alignment, demonstrating up to a 60% reduction in attack success rates and an over 84% decrease in unnecessary rejections. AI

IMPACT Highlights critical vulnerabilities in VLM safety training, potentially shifting alignment strategies towards more robust methods like machine unlearning.

RANK_REASON Academic paper detailing a new finding and proposed mitigation for VLM safety. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Machine Unlearning
• Vision-Language Models
• Yiwei Chen

• paper
• safety

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →