Engage the LLM in refining user questions to obtain more accurate and relevant answers, improving initial prompts.
