AI Wiki · Mediocre Security
Concept

Embedding

A numerical vector representing text or images so that similar meanings sit close together; the basis of search and RAG.

An embedding maps text (or images) to a vector so that similar meanings sit near each other, enabling semantic search, clustering and RAG. Perplexity’s pplx-embed-v2-late models (MIT, 0.6B and 9B variants sharing one embedding space) are a 2026 example of late-interaction embeddings, which compare token-level vectors for better precision. Embeddings are stored in a vector database.

Connections

Embedding and everything one link away. Drag to move, scroll to zoom, click a node to open it.Open in the full graph →

What links here

Company Perplexity

Concept Retrieval-augmented generation (RAG) · Vector database

Sources

  1. Wikipedia, “Word embedding”