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

Human Preference Alignment on PickScore (Comprehensive Metrics)

23.55PickScore (Task)

VGPO

21.646822.140922.63523.1291Dec 13, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.12
23.555.974.1812823.73.34
2025.12
23.416.154.1612423.563.33
2025.12
23.415.94.2313223.613.39
2025.12
23.315.924.2212823.533.38
2025.12
21.725.394.078722.343.07