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

Factuality Evaluation on HaluEval QA (500 items)

86.4MC1 Score

Qwen3-30B-A3B

-2.41620.64243.766.758May 18, 2026
Updated 15d ago

Evaluation Results

MethodLinks
2026.05
86.471.98
2026.05
8367.12
2026.05
79.470.4
2026.05
78.861.53
2026.05
76.463.68
2026.05
75.858.8
2026.05
75.659
2026.05
6759.32
2026.05
6152.3
2026.05
57.450.23
2026.05
55.447.81
2026.05
49.247.79
2026.05
48.243.99
2026.05
4744.13
2026.05
43.643.38
2026.05
41.239.13
2026.05
34.837.82
2026.05
24.826.97
2026.05
15.412.33
2026.05
14.411.29
2026.05
1412.06
2026.05
12.813.75
2026.05
8.47.15
2026.05
5.83.11
2026.05
5.48.06
2026.05
5.24.15
2026.05
54.67
2026.05
3.45.3
2026.05
2.42.62
2026.05
16.99