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

Multiple-choice Question Answering on MMLU-Pro Chem.

72.8Accuracy

DuoAttention

-2.91216.74436.456.056Oct 9, 2025
Updated 6d ago

Evaluation Results

MethodLinks
2025.10
72.8
2025.10
72.2
2025.10
70.8
2025.10
69
2025.10
65
2025.10
64.6
2025.10
62.6
2025.10
62.4
2025.10
56
2025.10
55
2025.10
51
2025.10
50.8
2025.10
49.8
2025.10
47.4
2025.10
44.8
2025.10
43.4
2025.10
41.2
2025.10
36
2025.10
32.8
2025.10
32.8
2025.10
24.8
2025.10
22.8
2025.10
21.6
2025.10
19.8
2025.10
18.8
2025.10
17.4
2025.10
9
2025.10
7.4
2025.10
7.2
2025.10
3
2025.10
2.8
2025.10
2.2
2025.10
1.4
2025.10
1
2025.10
0.8
2025.10
0.4
2025.10
0.4
2025.10
0.4
2025.10
0.2
2025.10
0