option to log llm response

[ChuckHend]

have had several requests to auto write vectorize.rag() calls to a table. This would provide option for retrieving previous responses instead of calling llm again.