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

Direct Preference Optimization on UltraFeedback

69.92Accuracy

FedPDPO

63.648865.276966.90568.5331Mar 20, 2026
Updated 27d ago

Evaluation Results

MethodLinks
2026.03
69.92
2026.03
68.02
2026.03
67.48
2026.03
67.25
2026.03
66.45
2026.03
65.54
2026.03
65.47
2026.03
65.45
2026.03
64.67
2026.03
63.98
2026.03
63.89