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

Factuality Evaluation on HaluEval Sum (500 items)

63MC1 Score

GPT-OSS-20B

-0.4416.0332.548.97May 18, 2026
Updated 2mo ago

Evaluation Results

MethodLinks
2026.05
6356.96
2026.05
50.649.84
2026.05
47.226.14
2026.05
4749.01
2026.05
44.648.2
2026.05
43.648.3
2026.05
4146.92
2026.05
33.844.34
2026.05
32.220.41
2026.05
32.242.48
2026.05
31.618.37
2026.05
31.441.14
2026.05
30.240.79
2026.05
3010.83
2026.05
28.841.41
2026.05
27.440.26
2026.05
26.639.33
2026.05
2641.11
2026.05
24.238.27
2026.05
196.19
2026.05
16.47.66
2026.05
15.46.51
2026.05
13.65.03
2026.05
10.83.16
2026.05
9.86.7
2026.05
8.83.81
2026.05
7.83.08
2026.05
6.81.29
2026.05
3.81.94
2026.05
20.4