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

Zero-shot Evaluation on 6 Downstream Tasks

80.05Average Accuracy

Dense

35.74647.24858.7570.252Feb 6, 2026Feb 16, 2026Feb 26, 2026Mar 9, 2026Mar 19, 2026Mar 29, 2026Apr 9, 2026
Updated 8d ago

Evaluation Results

MethodLinks
2026.02
80.05
2026.02
78.69
2026.02
75.81
2026.02
75.12
2026.02
74.89
2026.02
72.86
2026.02
72.72
2026.02
72.18
2026.02
71.85
2026.02
71.4
2026.02
69.23
2026.02
67.08
2026.02
67.05
2026.04
65.6
2026.04
64.9
2026.02
63.7
2026.02
63.03
2026.02
61.79
2026.04
60.6
2026.04
60.3
2026.02
59.71
2026.04
59.1
2026.04
59.1
2026.04
58.9
2026.04
58.5
2026.04
58.5
2026.04
57.3
2026.04
57.2
2026.04
56.3
2026.02
39.31
2026.02
37.45