Showing 2024Show all
2 papers · 1 filter
cs.CV2024
Explaining Caption-Image Interactions in CLIP Models with Second-Order Attributions
Lucas Möller, Pascal Tilli, Ngoc Thang Vu +1
Dual encoder architectures like Clip models map two types of inputs into a shared embedding space and predict similarities between them. Despite their wide application, it is, howe…
cs.CL2024
Approximate Attributions for Off-the-Shelf Siamese Transformers
Lucas Möller, Dmitry Nikolaev, Sebastian Padó
Siamese encoders such as sentence transformers are among the least understood deep models. Established attribution methods cannot tackle this model class since it compares two inpu…