{"url":"/dataset/tat-qa","name":"TAT-QA","full_name":null,"description_markdown":"TAT-QA (Tabular And Textual dataset for Question Answering) is a large-scale QA dataset, aiming to stimulate progress of QA research over more complex and realistic tabular and textual data, especially those requiring numerical reasoning.\r\n\r\nThe unique features of TAT-QA include:\r\n\r\n- The context given is hybrid, comprising a semi-structured table and at least two relevant paragraphs that describe, analyze or complement the table;\r\n- The questions are generated by the humans with rich financial knowledge, most are practical;\r\n- The answer forms are diverse, including single span, multiple spans and free-form;\r\n- To answer the questions, various numerical reasoning capabilities are usually required, including addition (+), subtraction (-), multiplication (x), division (/), counting, comparison, sorting, and their compositions;In addition to the ground-truth answers, the corresponding derivations and scale are also provided if any.\r\n\r\nIn total, TAT-QA contains 16,552 questions associated with 2,757 hybrid contexts from real-world financial reports.","description_withheld":null,"homepage":"https://nextplusplus.github.io/TAT-QA/","introduced_date":"2021-05-17","introduced_date_note":null,"introduced_by":{"paper":"/paper/tat-qa-a-question-answering-benchmark-on-a","title":"TAT-QA: A Question Answering Benchmark on a Hybrid of Tabular and Textual Content in Finance","first_author":"Fengbin Zhu","url":null},"license":{"name":"CC BY-NC-SA","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"modalities":[{"name":"Texts","url":"/datasets/modality/texts"},{"name":"Tables","url":"/datasets/modality/tables"}],"tasks":[{"name":"Question Answering","url":"/task/question-answering","datasets_with_task":"/datasets/task/question-answering"}],"languages":[{"name":"English","url":"/datasets/language/english"}],"variants":["TAT-QA"],"data_loaders":[],"num_papers_in_archive":76,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/question-answering-on-tat-qa","task":"Question Answering","dataset_variant":"TAT-QA","rows":1,"metrics":["Exact Match (EM)"],"first_row_in_archive_order":{"model":"TagOp","paper":"/paper/tat-qa-a-question-answering-benchmark-on-a","metrics":{"Exact Match (EM)":"50.1"},"code_links":[{"title":"NExTplusplus/TAT-QA","url":"https://github.com/NExTplusplus/TAT-QA"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/tat-qa-a-question-answering-benchmark-on-a","title":"TAT-QA: A Question Answering Benchmark on a Hybrid of Tabular and Textual Content in Finance","date":"2021-05-17","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":12,"samples_ran":2,"samples_unverified":10,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":1,"samples_harvested":12,"samples_ran":2,"samples_unverified":10,"pointer_only_for_licence":0,"papers_with_no_sample_that_ran":0,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}