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

Commonsense Reasoning on CMQA (Accuracy %)

70.47CMQA Accuracy (%)

Dense

37.34645.945554.54563.1445Oct 20, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.10
70.47
2025.10
69.99
2025.10
69.19
2025.10
67.61
2025.10
66.98
2025.10
66.68
2025.10
66.6
2025.10
66.3
2025.10
66.12
2025.10
65.87
2025.10
65.28
2025.10
65
2025.10
64.39
2025.10
63.48
2025.10
63.46
2025.10
63.34
2025.10
63.28
2025.10
63.27
2025.10
62.69
2025.10
62.64
2025.10
61.77
2025.10
61.27
2025.10
60.84
2025.10
60.84
2025.10
60.68
2025.10
59.95
2025.10
59.89
2025.10
59.66
2025.10
59.11
2025.10
58.54
2025.10
58.33
2025.10
58.3
2025.10
58.24
2025.10
57.75
2025.10
57.68
2025.10
57.5
2025.10
57.45
2025.10
57.3
2025.10
57.14
2025.10
56.9
2025.10
56.86
2025.10
56.51
2025.10
56.29
2025.10
55.75
2025.10
55.28
2025.10
54.36
2025.10
54.17
2025.10
53.51
2025.10
53.01
2025.10
52.24
2025.10
52.03
2025.10
51.89
2025.10
51.6
2025.10
51.5
2025.10
50.95
2025.10
50.12
2025.10
50.01
2025.10
49.79
2025.10
48.73
2025.10
48.54
2025.10
47.84
2025.10
47.46
2025.10
46.26
2025.10
46.09
2025.10
45.68
2025.10
45.35
2025.10
44.87
2025.10
44.82
2025.10
44.38
2025.10
43.61
2025.10
43.53
2025.10
43.52
2025.10
43.37
2025.10
41.99
2025.10
41.86
2025.10
41.82
2025.10
41.75
2025.10
41.75
2025.10
41.37
2025.10
41.32
2025.10
41.19
2025.10
41.1
2025.10
40.92
2025.10
39.68
2025.10
39.67
2025.10
39.13
2025.10
38.73
2025.10
38.62