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

Search-QA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Question AnsweringSearch-QA
Average Score51
177
Multi-turn Agent TaskSearch-QA
Success Rate50.49
31
Search-augmented QASearch-QA avg. of HotpotQA, 2Wiki, MuSiQue, and Bamboogle
Rollout Speedup Factor2.24
2
Showing 3 of 3 rows