Skip to main content
ModelTerms

Visualization

The AI vocabulary, drawn.

148 terms across 10 categories, 498 cross-links between related concepts. Each dot is a term; each curve is a connection between concepts that genuinely interact in practice.

FoundationsArchitectureTrainingInferencePromptingEvaluationSafety & AlignmentInfrastructureAgents & ToolsMultimodal

Hover a dot to highlight its connections · click to pin · click again to open · click a category badge to hide/show its terms

How to read it

Categories anchor the outer rim. Each dot inside a category is one term — bigger dots have more outgoing connections. Edges curve through the middle so the global graph stays readable even with ~500 of them.

Two kinds of edges appear: within-category (drawn in the category color, thicker) and cross-category (drawn in muted gray, thinner). The cross-category lines are where most of the conceptual interest sits — they show, for example, that RAG sits in Agents & Tools but reaches deep into Architecture, Evaluation, and Inference.

Prefer reading? Try the curated learning paths or browse the A–Z index.