AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Block Recurrent Dynamics in Vision Transformers

Mozes Jacobs, Thomas Fel, Richard Hakim, Alessandra Brondetta, Demba E. Ba, T. Anderson Keller

ICLR 2026 · Architectures and efficiency › Attention and transformers · approach: Hybrid and interpretable ViT designs · 9 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add autoregressive loss with annealed teacher forcing (ablation)Raptor k=3 · ImageNet-1ktop-1 accuracy: 3.9 -> 72.7 (+68.8)Improved
add depth scaling (learned layer-index embedding) (ablation)Raptor k=3 · ImageNet-1ktop-1 accuracy: 72.7 -> 75.2 (+2.5)Improved
up-weight cls token loss in final block (ablation)Raptor k=3 · ImageNet-1ktop-1 accuracy: 75.2 -> 76.7 (+1.5)Improved
add second stage end-to-end autoregressive training (ablation)Raptor k=3 · ImageNet-1ktop-1 accuracy: 76.7 -> 82.4 (+5.7)Improved
fine-tune the linear classifier (ablation)Raptor k=3 · ImageNet-1ktop-1 accuracy: 82.4 -> 83.0 (+0.6)Improved
increase recurrent blocks from k=2 to k=3Raptor · ImageNet-1ktop-1 accuracy: 81.2 -> 83.0Improved
increase recurrent blocks from k=3 to k=4Raptor · ImageNet-1ktop-1 accuracy: 83.0 -> 83.2No clear effect
increase recurrent blocks from k=2 to k=3Raptor · ADE20kmIoU: 39.6 -> 43.0Improved
increase recurrent blocks from k=2 to k=3Raptor · NYUv2RMSE (lower better): 0.648 -> 0.618Improved
increase stochastic depth probability during ViT training (ablation)ViT-B/14 from scratch · CIFAR-100layerwise similarity and Raptor reconstruction fidelityImproved
compare Raptor k=3 to DINOv2 ViT-SRaptor k=3 · ImageNet-1ktop-1 accuracy: 80.9 -> 83.0Improved
compare Raptor k=3 to DINOv2 ViT-BRaptor k=3 · ADE20kmIoU: 47.5 -> 43.0Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy