How ever with use and user inputs do the LLMs learn?
Would I be right in assuming that an Inference does not change and as a result does not learn?