AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

Concept-Guided Fine-Tuning: Steering ViTs away from Spurious Correlations to Improve Robustness

Yehonatan Elisha, Oren Barkan, Noam Koenigstein

CVPR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Parameter-efficient visual tuning · 1 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
concept-guided relevance fine-tuning (CFT)ViT-B · ImageNet-ATop-1: 13.26 -> 27.76Improved
CFT instead of RRDA baselineRRDA · ImageNet-ATop-1: 25.12 -> 27.76Improved
use concept masks instead of ground-truth object segmentation masks (ablation)CFT · ImageNet-ATop-1: 24.12 -> 26.05Improved
add classification-consistency loss (ablation)CFT · ImageNet-ATop-1: 17.96 -> 24.11Improved
add non-concept suppression loss (ablation)CFT · ImageNet-ATop-1: 19.14 -> 24.11Improved
add concept relevance loss (ablation)CFT · ImageNet-ATop-1: 24.10 -> 24.11No clear effect
use classification-consistency loss instead of ground-truth CE (ablation)CFT · ImageNet-ATop-1: 21.35 -> 24.11Improved
use AttnLRP instead of GradCAM relevance (ablation)CFT · ImageNet-ATop-1: 25.88 -> 27.82Improved
fine-tune on half of classes, evaluate held-out classes (ablation)CFT · ImageNet-A held-out classesTop-1: 17.20 -> 25.43Improved

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