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

Commonsense Reasoning on CommonsenseQA (pass@1 accuracy)

86.6Accuracy (pass@1)

DENOISE

38.55251.02663.575.974Mar 13, 2024Jul 20, 2024Nov 27, 2024Apr 6, 2025Aug 13, 2025Dec 21, 2025Apr 30, 2026
Updated 2mo ago

Evaluation Results

MethodLinks
2026.04
86.6
2026.04
86.2
2026.04
86
2026.04
86
2026.04
85.59
2026.04
85.59
2026.04
85.42
2025.10
85.34
2026.04
85.34
2026.04
85.26
2026.04
85.26
2026.04
85.18
2026.04
85.09
2026.04
85.09
2026.04
85.09
2026.04
84.93
2026.04
84.93
2026.04
84.8
2026.04
84.77
2026.04
84.68
2026.04
84.68
2026.04
84.6
2026.04
84.52
2026.04
84.52
2026.04
84.44
2026.04
84.44
2026.04
84.36
2026.04
84.36
2026.04
84.28
2026.04
84.19
2026.04
84.11
2026.04
84
2025.10
83.95
2025.10
83.95
2026.04
83.8
2025.10
83.7
2026.04
83.7
2026.04
83.54
2026.04
83.13
2026.04
83.1
2026.04
83
2026.04
82.8
2026.04
82.64
2026.04
82.56
2026.04
82.4
2026.04
82.31
2026.04
82
2026.04
81.8
2026.04
81.8
2026.04
81.8
2026.04
81.4
2026.04
81.3
2026.04
81
2026.04
81
2026.04
80.8
2026.04
80.7
2026.04
80.59
2026.04
79.52
2026.04
79.3
2026.04
79.3
2026.04
79.2
2026.04
78.5
2026.04
78.1
2026.04
76.9
2026.04
76.9
2026.04
74.7
2024.03
73.81
2024.03
71.33
2024.03
68.96
2026.04
68.9
2026.04
68.7
2026.04
63.4
2026.04
58.7
2026.04
55.8
2026.04
55.6
2026.04
55.6
2026.04
54.9
2026.04
53.4
2026.04
53.2
2026.04
52.3
2026.04
52.22
2026.04
51.6
2026.04
49.6
2026.04
48.4
2026.04
47.2
2026.04
46.8
2026.04
46.5
2026.04
46.3
2026.04
44.8
2026.04
44.4
2026.04
44.3
2026.04
43.4
2026.04
42.5
2026.04
42.3
2026.04
42.07
2026.04
41.66
2026.04
40.67
2026.04
40.46
2026.04
40.43
2026.04
40.4
Showing 100 of 108 rows