This dataset contains around 1M documents and 27K queries related to those documents.
1. `documents.jsonl` contains the documents. There are two fields `id` and `text`.
2. `queries.jsonl` contains the queries. Each query is composed of a question, an answer, a list of golden documents, and also a list of noisy documents.