HotpotQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
HotpotQA (test)
28.4EM
18
HotpotQA
100wAA
17
HotpotQA Poison (test)
48.8Retrieval Accuracy (RACC)
17
HotpotQA
52Success Rate (SR)
17
HotpotQA
65.7SubEM
17
HotpotQA (dev test)
71.2F1 Score
17
HotPotQA
28.1MBE
17
HotpotQA in-distribution
66.51Exact Match (EM)
17
HotpotQA (official evaluation)
33.2EM Accuracy
17
HotpotQA
0.574F1 Score
17
HotpotQA Modified (test)
42.04Accuracy
16
HotpotQA (test)
65.71Accuracy
16
HotpotQA
57Accuracy (Clean)
16
HotpotQA In-domain (test)
98.3Accuracy
16
HotpotQA (val)
42EM
16
HotpotQA fixed (test)
57.81EM
16
HotPotQA
55.5Initial Accuracy
16
HotpotQA
43.5Exact Match Accuracy
16
HotpotQA
42.4EM
16
HotpotQA
0.739AUROC
16
HotpotQA
98Precision
16
HotpotQA
100ATR
16
HotpotQA
84.2AUROC (128 Steps)
16
HotpotQA
0.681PR-AUC
16
HotpotQA
87.8P@1
16