Text embedding map

An exploratory semantic view of the corpus. Each point embeds the record's definition, canonical components, functions, taxonomic scope, and physical-property context.

The committed sentence-transformers/all-MiniLM-L6-v2 vectors are projected with PCA after unit normalization. Two-dimensional distance is lossy; use cosine neighbours for the full-vector comparison.

Select a structure

Choose a labelled point to inspect its nearest semantic neighbours.