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

NQ, TriviaQA, and WebQ

Benchmarks

Task NameDataset NameSOTA ResultTrend
Question AnsweringNQ, TriviaQA, and WebQ (test)
Accuracy65.5
21
Retrieval-Augmented Question AnsweringNQ, TriviaQA, and WebQ (test)
Answer Quality Recall62.2
18
Showing 2 of 2 rows