Research map / Architectures and efficiency / RNNs and normalization
How Well do Feature Visualizations Support Causal Understanding of CNN Activations?
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| show synthetic feature visualizations versus no reference images | Human occlusion prediction · Occlusion 2AFC task, MTurk | accuracy (%): 60 \pm 3 -> 67 \pm 4 | Helped |
| show synthetic feature visualizations versus other visualization types (natural, mixed, blurred) | Human occlusion prediction · Occlusion 2AFC task, MTurk | accuracy (%): 66 \pm 3 to 67 \pm 5 for others vs 67 \pm 4 | No clear effect |
| natural instead of synthetic references on the earlier maximally-activating-image choice task | Human unit-activation judgment · Reference replication, MTurk | accuracy (%): 65\pm3 -> 84\pm3 | Helped |
Closest papers by meaning
- Don't trust your eyes: on the (un)reliability of feature visualizations (ICML 2024 · similarity 0.85)
- Parallel Backpropagation for Shared-Feature Visualization (NeurIPS 2024 · similarity 0.85)
- Measuring Per-Unit Interpretability at Scale Without Humans (NeurIPS 2024 · similarity 0.84)
- Causal Tracing of Object Representations in Large Vision Language Models: Mechanistic Interpretability and Hallucination Mitigation (AAAI 2026 · similarity 0.84)
- Imagination Helps Visual Reasoning, But Not Yet in Latent Space (ICML 2026 · similarity 0.84)
- Adversarial Attacks on the Interpretation of Neuron Activation Maximization (AAAI 2024 · similarity 0.84)
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.