A model that converts text sentences into numerical vectors (embeddings) that capture their semantic meaning, enabling comparison of how similar different sentences are.
World knowledge accuracy, recall of facts and relationships