QA dataset
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Filtered QA dataset
97.65ASR (Naive)
30
QA Dataset 50 questions Long condition
72Accuracy
8
QA Dataset 100 questions - Short condition
94Accuracy
8
QA Dataset 150 questions Overall
86.7Accuracy
8
QA dataset Reverse direction
87Exact Match Accuracy
2
QA dataset Same direction
100Exact Match Accuracy
2