Hi Soumitra,Yes, this is a common issue with RAG systems. Even with the same context and prompt, LLMs can sometimes generate different responses.A few things I would check:Keep temperature low (0–0.2) for more consistent results.Make the prompt clear...
