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

Pairwise Preference Prediction on DyCoBench-1K Overall Preference

78.3Preference Rate (A > B)

DYCORM(OURS)

58.5463.6768.873.93May 25, 2026
Updated 8d ago

Evaluation Results

MethodLinks
2026.05
78.380.770.179.1
2026.05
72.874.468.773.2
2026.05
71.272.867.171.6
2026.05
70.872.466.771.2
2026.05
66.868.461.767.2
2026.05
65.567.161.465.9
2026.05
65.366.961.265.7
2026.05
65.166.76065.5
2026.05
64.766.361.365.1
2026.05
6465.657.964.4
2026.05
63.565.158.463.9
2026.05
63.264.858.163.6
2026.05
61.262.855.161.6
2026.05
60.46254.360.8
2026.05
59.861.453.760.2
2026.05
59.661.253.560
2026.05
59.360.953.259.7