Home▸AI & Machine Learning▸Word Embedding Space (2D)

Word Embedding Space (2D)

2D companion to the 3D word-embedding explorer: the same hand-placed coordinates and cosine-similarity math, projected onto a flat, pannable canvas — filter by semantic category, redraw neighbor links live and toggle the king − man + woman ≈ queen analogy vector.

AI & Machine Learning2DModerate60 FPS📱 Mobile-adapted⇄ 3D version

This 2D companion keeps the exact word coordinates and cosine-similarity math of the 3D word-embedding explorer, dropping the third axis so the whole space fits on a flat, pannable canvas: a category filter isolates one semantic cluster (animals, emotions, technology, royalty, food) at a time, a neighbor-link threshold slider redraws the nearest-neighbor connections in real time, and the classic king − man + woman ≈ queen analogy toggle draws the vector arithmetic directly between the points, so semantic relationships that are usually abstract become something you can drag, zoom and click through.

⚙ Under the hood

2D word-embedding explorer with cosine-similarity neighbor links, category filtering and a king − man + woman ≈ queen analogy vector, sharing coordinates with the 3D version.

machine learningnatural languageword embeddingsnlpcosine similarityvector space

2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install