Vision Transformers (ViTs) have become a dominant architecture in computer vision, yet producing stable and high-resolution attribution maps for these models remains challenging. Architectural components such as patch embeddings and attention routing often introduce structured artifacts in pixel-level explanations, causing many existing methods to rely on coarse patch-level attributions. We introduce DAVE \textit{(\underline{D}istribution-aware \underline{A}ttribution via \underline{V}iT Gradient D\underline{E}composition)}, a mathematically grounded attribution method for ViTs based on a structured decomposition of the input gradient. By exploiting architectural properties of ViTs, DAVE isolates locally equivariant and stable components of the effective input--output mapping. It separates these from architecture-induced artifacts and other sources of instability.
@article{arxiv.2602.06613,
title = {DAVE: Distribution-aware Attribution via ViT Gradient Decomposition},
author = {Adam Wróbel and Siddhartha Gairola and Jacek Tabor and Bernt Schiele and Bartosz Zieliński and Dawid Rymarczyk},
journal= {arXiv preprint arXiv:2602.06613},
year = {2026}
}
Comments
work under review. Code will be released upon acceptance