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

Best-of-N Reward Evaluation on UltraFeedback core250

24.323Reward Score

TEA

11.1326814.5570917.981521.40591May 11, 2026
Updated 21d ago

Evaluation Results

MethodLinks
2026.05
24.3231.102
2026.05
23.3421.067
2026.05
23.221-
2026.05
22.275-
2026.05
22.090.835
2026.05
21.255-
2026.05
20.8580.701
2026.05
20.157-
2026.05
19.5220.589
2026.05
18.933-
2026.05
18.0140.485
2026.05
17.529-
2026.05
16.3220.383
2026.05
15.939-
2026.05
14.4930.469
2026.05
14.024-
2026.05
12.0810.441
2026.05
11.64-