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

Anthropic Helpful

Benchmarks

Task NameDataset NameSOTA ResultTrend
Reward Model TransferAnthropic Helpful (AH)
AOG9.95
16
Reward ModelingAnthropic Helpful
Accuracy72
12
Open-ended DialogueAnthropic-Helpful (ID)
Win Rate0.762
4
Showing 3 of 3 rows