Looking forward to seeing where your synthetic dataset generation goes! Grounding results with RAG does seem like a great approach for creating data that is hopefully both correct and varied
thanks for the reply, Logan. I agree, I was chomping at the bit to add llama index and see how much extra mileage grounding could give, and was pleasantly surprised.
