HotpotQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
HotpotQA
57.23NDCG@10
6
HotpotQA 5Q
15.4Exact Match Accuracy
6
HotpotQA 3Q
32Exact Match Accuracy (3Q)
6
HotpotQA
100Answerability Rate
6
HotpotQA
60.78F1 Score
6
HotpotQA In-Distribution
59.4Accuracy (8K Context)
6
HotpotQA
81Hit Rate
6
HotpotQA
70.3F1 (Token Overlap)
6
HotpotQA pilot
630Claims Emitted
6
HotPotQA (test)
56.34Accuracy
6
HotpotQA
33.5pass@1
6
HotpotQA (dev)
59.8Accuracy
6
HotpotQA (test)
33.46F1 Score
6
HotpotQA sampled 1,000 questions (val)
44.5Swedish EM
6
HotpotQA (test val)
59.8EM
6
HotpotQA
62.5F1
6
HotpotQA
49.6Exact Match (EM)
6
HotpotQA
0.63Avg Running Time (s)
6
HotpotQA
84.6Accuracy
6
HotpotQA
83Accuracy
6
HotpotQA
0.278Inference Cost (USD)
6
HotpotQA first 1000 samples (dev)
39.7EM
6
HotpotQA (test)
48.72F1 Score
6
HotpotQA
-0.015Unattacked Avg |PS|
6
HotpotQA
87.3Reverse Accuracy
6