> 4) Use a local LLM embedding model to convert words to vectors, then cluster themOh my, I didn't expect that such care was taken to pick the words. That's a fun use of an embedding! Kudos!
