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

Cloze Task on MLLMU-Bench (Retain Set)

24.52Accuracy

Noise-T

0.06966.417312.76519.1127May 26, 2026
Updated 1d ago

Evaluation Results

MethodLinks
2026.05
24.52
2026.05
24
2026.05
23.15
2026.05
20
2026.05
19.13
2026.05
19.12
2026.05
19
2026.05
18.81
2026.05
18
2026.05
17.34
2026.05
16.16
2026.05
16.16
2026.05
16.16
2026.05
16.16
2026.05
15.66
2026.05
15.66
2026.05
15
2026.05
14.65
2026.05
11.63
2026.05
11.62
2026.05
11.21
2026.05
10.78
2026.05
10.57
2026.05
10.15
2026.05
9.19
2026.05
9
2026.05
5.5
2026.05
1.51
2026.05
1.26
2026.05
1.01