Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Commonsense Question Answering on CommonsenseQA blind v1.0 (test)

75.3Accuracy

Our Model

55.95660.9786671.022Sep 9, 2019
Updated 1mo ago

Evaluation Results

MethodLinks
2019.09
75.3
2019.09
72.5
2019.09
72.1
2019.09
72.1
2019.09
69.6
2019.09
68.4
2019.09
66.9
2019.09
65.3
2019.09
64.6
2019.09
62.9
2019.09
62.5
2019.09
62.5
2019.09
62.2
2019.09
61.8
2019.09
59.6
2019.09
58.9
2019.09
58.2
2019.09
57.9
2019.09
57.1
2019.09
56.7