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

Multi-task Language Understanding on MMLU-PRO (Acc %, ECE)

51.91Accuracy (%)

gemma3-27b-slerp

39.336442.600745.86549.1293Oct 20, 2025
Updated 22d ago

Evaluation Results

MethodLinks
2025.10
51.910.219
2025.10
51.460.192
2025.10
51.440.156
2025.10
51.110.254
2025.10
50.960.284
2025.10
50.890.119
2025.10
50.840.122
2025.10
50.580.312
2025.10
50.410.077
2025.10
50.230.078
2025.10
49.780.341
2025.10
49.390.037
2025.10
47.80.478
2025.10
44.250.219
2025.10
44.170.221
2025.10
44.120.173
2025.10
44.110.173
2025.10
44.070.265
2025.10
44.040.266
2025.10
440.122
2025.10
43.970.126
2025.10
43.940.125
2025.10
43.820.312
2025.10
43.790.232
2025.10
43.760.311
2025.10
43.640.179
2025.10
43.360.075
2025.10
43.320.302
2025.10
43.260.282
2025.10
43.240.353
2025.10
43.220.352
2025.10
43.190.079
2025.10
43.030.078
2025.10
42.50.337
2025.10
42.350.024
2025.10
42.150.395
2025.10
42.130.395
2025.10
41.520.406
2025.10
41.430.43
2025.10
41.360.431
2025.10
40.430.448
2025.10
39.820.533