HellaSwag
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
HellaSwag 10-shot
53.41HellaSwag 10-shot Accuracy
9
HellaSwag 200-sample
67Accuracy
8
HellaSwag
3.23ECE
8
Hellaswag
78.5Score
8
HellaSwag
50.95Accuracy
8
HellaSwag (HS)
63.1Accuracy
8
HellaSwag
60Accuracy
8
HellaSwag 10-shot
81.7Normalized Log Accuracy
8
HellaSwag EL
52.58Accuracy
8
HellaSwag
40.99Score
8
Hellaswag
60.6Accuracy (Before Attack)
8
HellaSwag Spanish
54.4Normalized Accuracy
7
HellaSwag
0.007MAE
7
HellaSwag
88EM
7
HellaSwag 10-shot
92.12Score
7
HellaSwag
54.6Accuracy (EN)
6
HellaSwag
68Accuracy (EN)
6
HellaSWAG
63.84HellaSWAG Normalized Accuracy
6
Hellaswag
64.2Accuracy
6
Hellaswag EU20 (test)
48.83Accuracy
6
Hellaswag de
49.28Accuracy
6
Hellaswag French
54.36Score
6
HellaSwag 2K (test)
76.8Accuracy
6
HellaSwag 2K Synonym Replacement WordNet-based (test)
65.4Accuracy
6
Hellaswag 2K (test)
17.4Failure Rate
6