A classification-trained ViT encodes patch boundaries at layers 5-6 and depth at layer 8, with causal interventions showing the depth signal is actively re-derived rather than passively carried.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
From Edges to Depth: Probing the Spatial Hierarchy in Vision Transformers
A classification-trained ViT encodes patch boundaries at layers 5-6 and depth at layer 8, with causal interventions showing the depth signal is actively re-derived rather than passively carried.