Hello, there LlamaIndex discord!! What are your approaches to answer "enumerate questions" from documents. if each example or part of the answer was in a different part of the document. How can the retriever and the LLM collaborate together to achieve the final solution?
Agents already can answer multihop questions, yet fall short in answering "enumerate questions"..
