Home▸AI & Machine Learning▸AI Interpretability Lab: Feature Attribution, Saliency & Attention Explorer

🔍 AI Interpretability Lab: Feature Attribution, Saliency & Attention Explorer

Explore how a small neural network explains its own decisions — switch between global feature importance, local gradient x input saliency and animated attention flow across a live 6-8-8-2 network, and watch the attribution bars and explanation entropy update as you perturb the input.

AI & Machine Learning3DModerate60 FPS

A live 6→8→8→2 neural network explains its own credit-approval decision in real time — switch between global feature importance, local gradient × input saliency, and an animated attention-style flow, then perturb the input to see the explanation change.

🔬 What It Demonstrates

Model interpretability is not one technique but a family of them. This simulation runs a real (small) neural network with an actual backpropagation pass, so the "Saliency" mode is a genuine gradient × input attribution — not a canned animation — while "Feature Importance" shows the simpler, global weight-based view that ignores the current input entirely. Watching both side by side makes the local-vs-global distinction concrete.

🎮 How to Use

Pick an explanation method, choose a feature to perturb and drag its value slider, or hit "New sample" for a fresh random input. Switch to "Attention Flow" to watch particles travel from the most-attributed input feature, through the hidden layers, to the predicted class. Drag to orbit the network and scroll to zoom.

💡 Did You Know?

"Gradient × Input" and its more robust cousins (Integrated Gradients, SHAP, LIME) are exactly how real interpretability tools explain production models — from credit scoring to medical diagnosis — and are increasingly required by regulations such as the EU's GDPR "right to explanation".

⚙ Under the hood

Explore how a small neural network explains its own decisions — switch between global feature importance, local gradient x input saliency, and animated attention flow across a live 6-8-8-2 network, and watch the attribution bars and explanation entropy update as you perturb the input.

Three.jsExplainable AIFeature AttributionSaliency MapsAttentionSHAPInstancedMesh

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install