ShieldCLIP: Selective Safety Alignment for Harmful Content Mitigation in Multimodal Foundation Models
Multimodal encoders such as CLIP underlie many downstream systems, but their web-scale training data embed harmful associations that safety alignment must suppress without unnecessarily changing benign representations. Because ethical and practical constraints prevent collecting real unsafe content at scale, existing d...
Tobia Poppi, Silvia Cappelletti, Samuele Poppi et al.
· 0 citations