Semantic Map
Every verse encoded as a 384-dimensional vector, then projected to 2D via UMAP. Proximity in this space reflects semantic similarity — not canonical order.
Loading semantic map
Selected Verse
Click any point on the map to see the verse and its semantic neighbors.
What you're seeing
Each dot is one of ~31,000 KJV verses. The position is computed by encoding the verse text with a sentence transformer (all-MiniLM-L6-v2) to produce a 384-dimensional embedding vector, then reducing to 2D using UMAP (n_neighbors=15, min_dist=0.1, cosine metric). Points that are close together have similar semantic content — even if they come from different books, testaments, or genres.
Try this
- Switch color to "Book (sequential)" and look for books that are scattered vs. tightly clustered — high scatter suggests thematic breadth.
- Filter to "Wisdom" genre and observe how Proverbs, Ecclesiastes, and Job occupy different regions despite being in the same genre.
- Click a verse in a dense cluster and check its neighbors — are they from the same book, or do they cross canonical boundaries?