multi-type answer

answer format (1) retriever-reader model (1) end-to-end closed-book qa (1) natural question (1) large language model (1) question answering benchmark (1) answer type (1) open-domain question answering (1) benchmark evaluation (1)