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

Model Merging on 7-task NLP benchmark

1.18Avg Performance

METIS

0.494640.672570.85051.02843Jun 15, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
1.180.92-0.26
2026.06
1.1080.886-0.22
2026.06
1.0670.75-0.32
2026.06
1.0480.75-0.3
2026.06
1.0150.872-0.14
2026.06
0.9860.614-0.37
2026.06
0.9420.641-0.3
2026.06
0.9350.585-0.35
2026.06
0.8830.375-0.51
2026.06
0.8520.39-0.46
2026.06
0.8380.634-0.2
2026.06
0.8070.542-0.27
2026.06
0.80.525-0.28
2026.06
0.7520.4-0.35
2026.06
0.7260.275-0.45
2026.06
0.7060.385-0.32
2026.06
0.7040.39-0.31
2026.06
0.6940.293-0.4
2026.06
0.6630.375-0.29
2026.06
0.5210.08-0.44