AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

Specialization after Generalization: Towards Understanding Test-Time Training in Foundation Models

Jonas Hübotter, Patrik Wolf, Aleksandr Shevchenko, Dennis Jüni, Andreas Krause, Gil Kur

ICLR 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Pseudo-label test-time adaptation

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 test-time training (local last-layer fine-tuning on 50 neighbors)global linear classifier · ImageNet-1k (CLIP ViT-B/32)accuracy: 78.33 -> 79.39Improved
add test-time training (local last-layer fine-tuning)LeNet-5 · MNISTaccuracy: 98.57 -> 99.01Improved
add TTT on SAE concept featuresglobal linear on SAE features · ImageNet-1kaccuracy: 71.45 -> 72.64Improved
increase model size (ablation)TTT vs global gap · MNIST, ImageNet, Pileerror / bits per byteMixed
increase training set size (ablation)TTT vs global gap · MNIST and ImageNeterror rateImproved
apply a single local TTT head globally (ablation)TTT heads · MNISTaccuracy: 98.57 -> 36.38Hurt
apply a single local TTT head globally (ablation)TTT heads · ImageNet-1kaccuracy: 78.33 -> 77.04Hurt
use non-adaptive mask instead of learned sparse mask (ablation)sparse TTT · ImageNet-1k (SAE)accuracy: 72.56 -> 71.51Hurt
add learned sparse concept mask (~40 concepts) (ablation)TTT on SAE reconstructions · ImageNet-1kaccuracyNo clear effect
use TTT end-to-end instead of last layer only (ablation)MNIST classifier · MNISTaccuracyNo clear effect
compare TTT to majority vote over neighborhood (ablation)TTT · image classification with many classeserrorImproved

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