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

OQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Question AnsweringOQA
Accuracy49.8
24
Reward-wise QA fairness and alignmentOQA
JS Divergence (FI)0.927
15
Ordinal Preference AlignmentOQA
FI Bor.72.13
15
Opinion AlignmentOQA
Opinion Alignment91.6
8
Showing 4 of 4 rows