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

RubricBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Pairwise AccuracyRubricBench 1.0 (test)
Pairwise Accuracy (Chat)81.8
13
Reward ModelingRubricBench
Instruction Following (IF) Score59.7
8
Showing 2 of 2 rows