data
Graph-structural embeddings for corpus entities — methods, datasets, benchmarks, labs. Used for similarity and clustering.
Node Embedding produces vectors for the entity graph: methods, datasets, benchmarks, labs, opportunity clusters. The embeddings power similar-method discovery, cluster summaries, and the Talent search modes. They sit alongside Paper Chunker output but live one layer up: the chunks describe what a paper says, the node embeddings describe what entities the corpus knows.
Source: curated glossary catalog. Freshness: git_versioned_curated_catalog.
Term API · apps/web/data/glossary/terms.ts