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

Human Preference Alignment on Out-of-Domain (test)

81.8Agreement

Annotators

23.87238.91153.9568.989Mar 25, 2026
Updated 23d ago

Evaluation Results

MethodLinks
2026.03
81.8--
80.27773
2026.03
59.54140
2026.03
58.844.534
2026.03
57.63834.5
2026.03
56.93237.5
2026.03
56.43633.5
2026.03
54.340.532
2026.03
49.738.129.9
2026.03
47.942.636.6
2026.03
44.150.539
2026.03
43.32619
2026.03
40.738.536
2026.03
35.525.58.5
2026.03
26.15957