AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Disentangling Hallucinations: Orthogonal Semantic Projection for Robust Interpretability

Emirhan Bilgiç, Baptiste Caramiaux, Zhi Yan, Gianni Franchi

ECCV 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded visual concept learning · 0 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 OSPAttentionCAM on CLIP · ImageNet-SegmentationAUROC: 52.68 -> 67.73Improved
add OSPCheferCAM on SigLIP · ImageNet-SegmentationAUROC: 55.47 -> 62.64Improved
add OSPGradCAM on CLIP · ImageNet-SegmentationAUROC: 65.39 -> 71.82Improved
add OSPLeGrad on SigLIP · ImageNet-SegmentationAUROC: 74.42 -> 74.73Improved
add OSPDAAM on Stable Diffusion 2 · ImageNet-SegmentationAUROC: 83.07 -> 85.48Improved
add OSPAttentionCAM on CLIP · ImageNet-Segmentation positive promptsmIoU: 40.14 -> 47.96Improved
add OSPLeGrad on CLIP · ImageNet-Segmentation negative promptsmIoU: 40.88 -> 40.74No clear effect
add OSPCheferCAM on SigLIP · ImageNet-Segmentation negative promptsmIoU: 36.65 -> 38.45Hurt
add OSPLeGrad on SigLIP · ImageNet-Segmentation negative promptsmIoU: 37.27 -> 34.68Improved
use ImageNet-class dictionary instead of no OSP (ablation)OSP, avg 9 pairs · ImageNet-SegmentationAUROC: 70.64 -> 74.08Improved
use WordNet-augmented dictionary (ablation)OSP, avg 9 pairs · ImageNet-SegmentationAUROC: 70.43 -> 73.90Improved
use Gemini 3 Flash dictionary (ablation)OSP, avg 9 pairs · ImageNet-SegmentationAUROC: 70.65 -> 75.52Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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