Multi-hop QA Benchmarks
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Multi-hop QA Benchmarks (2WikiMultiHopQA, Bamboo, HotpotQA, MuSiQue, Natural Questions, PopQA, TriviaQA)
79.1Accuracy (2WikiMultiHopQA)
17
Multi-hop QA Benchmarks (NQ, Trivia, Bamboogle, HotpotQA, 2WikiMultihopQA, MuSiQue)
78NQ Accuracy
7