cs.CVNov 17, 2026

Keeping Parts Honest: Isolated Part Graphs for Faithful Vision Transformer Explanations

Authors: Bernard Tomczyk, Gail Collyer-Hoar, Bryan Williams

Organizations: Lancaster University, Lancaster, United Kingdom

Abstract

Part and prototype-based explanations for Vision Transformers (ViTs) present a decision as evidence from individual image regions. However, self-attention mixes information across patches, so part-level features absorb context from the whole image and consequently fail to faithfully reflect the decision. We show that existing interpretable ViT classifiers depend heavily on this leakage: when each explained region is encoded in isolation, their accuracy collapses, and retraining them under isolation recovers only part of the lost accuracy. We propose a classifier that represents each image as a graph of isolated parts. Each part is encoded only from its own patches by a ViT whose last blocks are fine-tuned under isolation, preventing evidence leakage from elsewhere in the image. Parts are discovered without annotation and refined hierarchically through their sub-parts, making them more consistent and stable across images than those of competing methods. Each edge is computed from part pairs, and recovers evidence that attention previously shared across the object without affecting the parts. Matching the image graph to a learned prototype graph per class makes the prediction a direct combination of node and edge similarities, each computed only from the patches it is credited to, so every explanation is faithful by construction. Across 11 fine-grained and general-object benchmarks, our method exceeds the accuracy of existing explainable methods that rely on leakage, while providing explanations that reflect what the model actually uses.

Figures & tables

Explore similar work

CardsList
  1. Decision-Aware Attention Propagation for Vision Transformer Explainability

    Apr 20, 2026Sehyeong Jo, Gangjae Jang, Haesol ParkTransformer InterpretabilityVision Transformer

  2. The Explainability Illusion: Must ViTs Exhibit Patch-Decoding Interpretability?

    Oct 18, 2026Ronen Schaffer, Hila Manor, Tomer MichaeliTransformer InterpretabilityExplainable Image Classification

  3. Metonymy in vision models undermines attention-based interpretability

    May 7, 2026Ananthu Aniraj, Cassio F. Dantas, Dino Ienco +2Transformer InterpretabilityDisentangled Representation Learning