Embeddings
Words as points in space: click any word to find its nearest neighbours by real cosine similarity, and run king − man + woman = queen as genuine vector arithmetic.
Before a model can compute with words it turns each token into a vector — a list of numbers — positioned so that similar meanings sit close together. Distance in this space is meaning.
This workspace shows a 48-word vocabulary authored in 8 interpretable dimensions (animacy, royalty, gender, techiness, edibility, valence, nature, size), projected to 2D for display — every embedding scatter you have ever seen is a projection, including this one. Neighbour lookups are real cosine similarity; the analogy demo (king − man + woman ≈ queen) is real vector arithmetic over those 8 dimensions.