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

Multi-task Language Understanding on MMLU (MMLU Acc)

82.67MMLU Accuracy

CascadeDebate

-5.178817.628140.43563.2419Jun 16, 2025Aug 5, 2025Sep 24, 2025Nov 14, 2025Jan 3, 2026Feb 22, 2026Apr 14, 2026
Updated 3d ago

Evaluation Results

MethodLinks
2026.04
82.67
2026.04
78
2026.04
73.2
2026.04
67.7
2026.04
67.11
2026.04
65
2026.04
64.78
2026.04
62.44
2026.04
58.78
2026.04
58.44
2026.04
54.5
2026.04
54.5
2026.04
51.8
2025.06
48.6
2026.04
47.56
2025.06
45.4
2025.06
45.2
2026.04
44.22
2025.07
35.03
2025.07
34.65
2025.07
34.5
2025.07
34.47
2025.07
34.27
2025.07
34.22
2025.07
34.06
2025.07
34.06
2025.07
33.81
2025.07
33.78
2025.07
33.2
2025.07
32.83
2025.07
31.27
2025.06
31.2
2025.07
30.74
2025.07
30.66
2025.07
30.42
2026.04
30.1
2025.07
30
2025.06
29.7
2025.07
29.37
2025.07
29
2026.04
28.3
2025.07
27.25
2025.07
27.11
2025.06
27.1
2025.07
27.06
2025.06
26.8
2025.06
26.4
2025.06
26
2026.04
25.9
2025.07
25.8
2025.06
22.5
2025.06
21.9
2025.06
21.5
2026.04
3.1
2026.04
0.9
2026.04
0.1
2026.04
0.1
2026.04
0
2026.04
-1.8