You won’t get grounded results by asking for feedback. The LLM doesn’t know its own internal state so while you may get a plausible and contextually significant answer, it isn’t going to be an accurate one.