List your company

NQ, TriviaQA, PopQA, HotpotQA, 2Wiki, MuSiQue, Bamboogle

Benchmarks

TaskDatasetSOTA resultTrendResults
NQ, TriviaQA, PopQA, HotpotQA, 2wiki, MuSiQue, Bamboogle
48.2Average QA Score
55
NQ, TriviaQA, PopQA, HotpotQA, 2wiki, MuSiQue, Bamboogle (test)
39.3NQ Score
5
NQ, TriviaQA, PopQA, HotpotQA, 2Wiki, MuSiQue, Bamboogle (held-out val)
48.4Exact Match (EM)
5