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

Helpsteer2

Benchmarks

Task NameDataset NameSOTA ResultTrend
Attribute-controlled Text GenerationHelpSteer2 Negative Representative Target Score (test)
Diversity0.987
12
Instruction FollowingHelpsteer2 Trivial
Accuracy78.22
8
Showing 2 of 2 rows