A technique that groups similar texts together automatically by using embeddings to measure similarity, without requiring predefined categories.
World knowledge accuracy, recall of facts and relationships