Question Answering with Abstention
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
47BAS
24
May 12, 2026
0.2BAS
24
May 12, 2026
91.5MMLU Accuracy
24
Apr 3, 2026
91.4U-Ref
7
Apr 21, 2026
83.4Accuracy (A)
7
Apr 21, 2026
78.6Accuracy
7
Apr 21, 2026