TriviaQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
TriviaQA
47BAS
24
TriviaQA n=1000 (test)
0.794AUROC
24
TriviaQA
75.4EM Accuracy
24
TriviaQA (test val)
68.2EM
24
TriviaQA (test)
100Power (α=0.1)
24
TriviaQA (standard)
71.2Accuracy
24
TriviaQA TQ (test)
76.12EM
24
TriviaQA Inj
0.22FPR (None)
24
TriviaQA
74.96HSPP
23
TriviaQA
70.3Accuracy
23
TriviaQA
85.2Accuracy
22
TriviaQA
66.5PRR
22
TriviaQA
57.3Accuracy
22
TriviaQA filtered (dev)
0.73Exact Match
22
TriviaQA MRQA
78.2F1 Score
22
TriviaQA
75.97str-EM Accuracy (@5)
21
TriviaQA
61.8Success Rate
21
TriviaQA
68.5Exact Match (EM)
21
Inj-TriviaQA
0.11Naive ASR
21
TriviaQA
5.04ECE@1
21
TriviaQA
49.2Factual Accuracy
21
TriviaQA
87.48Faith
21
TriviaQA Wiki domain, Verified (dev)
63.8EM
21
TriviaQA
66.4Exact Match (EM)
20
TriviaQA
73.33Accuracy
20