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

Natural Language Understanding on SuperGLUE (CB, COPA, BoolQ, WiC subset)

94.5CB Accuracy

ARMADA

72.24478.02283.889.578Mar 11, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.03
94.581.987.357.6
2026.03
93.479.987.960
2026.03
9379.987.859.9
2026.03
92.881.988.159.6
2026.03
92.881.986.259.2
2026.03
92.878.987.758.2
2026.03
87.462.972.358.2
2026.03
87.46875.358.4
2026.03
85.767.775.858.5
2026.03
85.767.975.158.8
2026.03
85.664.973.356.2
2026.03
83.860.97358.7
2026.03
83.762.975.158.4
2026.03
83.762.975.158.4
2026.03
81.983.787.260.8
2026.03
81.983.787.260.8
2026.03
80.254.975.355.9
2026.03
78.6637357.9
2026.03
78.558.975.657
2026.03
78.56373.158.1
2026.03
78.557.977.356.7
2026.03
78.451.973.355.9
2026.03
78.254.874.657.9
2026.03
78.254.874.657.9
2026.03
76.862.873.157.8
2026.03
76.7577657
2026.03
76.758.976.156.1
2026.03
76.658.97657
2026.03
755371.355.7
2026.03
7562.972.458.2
2026.03
755371.355.7
2026.03
73.149.971.856.6