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

Preference Alignment on UltraFeedback

81Win Rate

CLIPer (Vanilla Baseline)

44.654.0563.572.95Oct 10, 2025Nov 14, 2025Dec 19, 2025Jan 23, 2026Feb 27, 2026Apr 3, 2026May 8, 2026
Updated 21d ago

Evaluation Results

MethodLinks
2026.05
81-
2026.05
79-
2026.05
75.33-
2025.10
67-3.12
2025.10
64-3.75
2025.10
63-3.59
2026.05
58.33-
2025.10
58-4.42
2025.10
56-9.25
2026.05
53.67-
2025.10
52-9.51
2025.10
52-9.7
2025.10
51-9.87
2025.10
50-5.47
2025.10
50-10.02
2026.05
46-