Embedding Explorer

Navigate the semantic space where meaning becomes geometry. Pan, zoom, and discover what's near what.

1.0x
Click and drag to pan. Scroll to zoom.
Nearest neighbors

    Try these clusters:

    What am I looking at?

    This is a 2D projection of an embedding space. In AI, embeddings are vectors — lists of numbers — that represent the meaning of words, concepts, or any kind of data. When you plot these vectors, things that are semantically similar end up close together.

    The ~200 points here span four categories: music genres, emotions, tech concepts, and fun/experiential items. Their positions are designed to reflect real semantic relationships:

    • Electronic, house, and techno cluster tightly because they share sonic DNA.
    • Jazz, blues, and soul sit near each other — and near positive emotions like joy and contentment.
    • Metal, punk, and grindcore cluster far from ambient and classical.
    • "Meditation" lives near "serenity" and "peace" across category boundaries.
    • ML/AI terms (transformer, attention, LLM) form their own dense cluster, separate from web/infra tech.

    In real embedding models (like those powering LLMs), these spaces have hundreds or thousands of dimensions. This 2D view is a simplified illustration — but the clustering principle is exactly how modern AI understands meaning.