Embeddings are the foundation of vector search, allowing us to represent meaning-rich content like documents or queries as numerical vectors. But to use them effectively, it’s essential to understand what’s actually being embedded—whether that’s individual words, full sentences, or larger chunks of text.
