English
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
English en
54.1QA-F1
24
English
98.8BPA
24
English (test)
0.017WER
21
English user-disjoint
63.6Main Score
20
English lm-evaluation-harness
0.819ARC Easy Acc (Norm)
16
English (en) (test)
95.33LAS
16
English
0.925Median End-to-end CPU Time (s)
15
English SPMRL (test)
69.7S-F1
15
English
13.35WER
12
English
3.13Speedup Factor
12
English (test)
62Precision
12
English (test)
72.7Overall Score
11
English eng
1.27NSL Score
10
English 300 MB Corpora
64.4Morph. Score
9
English
98.78Bias-Penalized Accuracy (BPA)
9
English
3.65Bytes per Token
6
English (test)
2.512LSD
6
English
88.2Span Score
6
English (en)
32WER
5
English high-resource
4.61MOS
5
English (test)
0.01MSE
5
English
96.57Accuracy
5
English subset
96.73Accuracy
5
English
97.04Accuracy
5
English
0.768Absolute Kendall's Tau
5