AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

What is Missing? Explaining Neurons Activated by Absent Concepts

Robin Hesse, Simone Schaub-Meyer, Janina Hesse, Bernt Schiele, Stefan Roth

ICML 2026 · Probabilistic and scientific modeling › Causal discovery · approach: Causal representation identifiability · 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 non-target (absence) attribution prior to presence debiasing (ablation)attribution-prior debiasing, X-ResNet-50 · synthetic ISIC, inverse-bias validationavg accuracy: 0.41 -> 0.59Improved
add non-target attribution prior (ablation)attribution-prior debiasing · synthetic ISIC, inverse-bias validationmalignant accuracy: 0.17 -> 0.43Improved
add non-target attribution prior (ablation)attribution-prior debiasing · synthetic ISIC, no-bias validationavg accuracy: 0.77 -> 0.81Improved
add non-target attribution prior (ablation)attribution-prior debiasing · synthetic ISIC, training-bias validationavg accuracy: 0.92 -> 0.89Hurt
apply any attribution-prior debiasing instead of none (ablation)X-ResNet-50 · synthetic ISIC, no-bias validationavg accuracy: 0.51 -> 0.77Improved
insert least-activating patch instead of random patchVGG19 · ImageNet-1k, patch 48, 100 imagesmean activation: 2.68 -> 0.94Improved
insert least-activating patch instead of random patchResNet-50 · ImageNet-1k, patch 48, 100 imagesmean activation: 0.16 -> 0.03Improved
insert most-activating patch instead of random patchVGG19 · ImageNet-1k, patch 48, 100 imagesmean activation: 2.68 -> 3.88Hurt
insert prior-work logical NOT patch instead of random patchVGG19 · ImageNet-1k, patch 48, 100 imagesmean activation: 2.68 -> 2.51No clear effect
increase patch size from 32 to 64 for least-activating patch (ablation)VGG19 · ImageNet-1k, 100 imagesmean activation: 2.14 -> -0.38Improved
measure least-activating patch in earlier layer layer3[2].conv3 vs layer4[2].conv3 (ablation)ResNet-50 · ImageNet-1kmean activation drop: 0.26 -> 0.18Mixed

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