Long Sequence
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Long Sequence Order 2
84.86Average Performance (AP)
20
Long Sequence (Order 1)
82.87AP
20
Long Sequence (test)
82.02AP
15
Long Sequence
60.92EM
12
Long Sequence
1.85FWT (Order 1)
9
Long Sequence
131.05Throughput
8
Long sequence
1.36Accuracy
6
Long Sequence
82.32AP
6
Long Sequence Order 2 Qwen-3-4B
83.32Average Performance (AP)
4
Long Sequence Order 1 Qwen-3-4B
83.98AP
4
Long sequence
6.88Sync-C
4