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

CLINC150

Benchmarks

Task NameDataset NameSOTA ResultTrend
Intent ClassificationClinc150 cross-domain
Average Accuracy93.5
38
Short-text ClusteringClinc150 (test)
NMI95.72
23
Intent ClassificationCLINC150
Accuracy98.45
17
Intent ClassificationCLINC150 DialoGLUE 10-shot
Accuracy95.58
9
Intent ClassificationClinc150 5-shot (test)
Accuracy89.69
8
Out-of-Domain DetectionCLINC150 75% known ratio (test)
Accuracy0.877
6
Out-of-Domain DetectionCLINC150 50% known ratio (test)
Accuracy88.61
6
Out-of-Domain DetectionCLINC150 25% known ratio (test)
Accuracy89.79
6
Unknown Intent DetectionCLINC150 75% seen classes (test)
Accuracy88.08
6
Unknown Intent DetectionCLINC150 50% seen classes (test)
Accuracy88.33
6
Unknown Intent DetectionCLINC150 25% seen classes (test)
Accuracy88.44
6
CalibrationClinc150
ECE0.236
4
Intent ClassificationCLINC150 (deployment)
Count of Failures Fixed453
2
Showing 13 of 13 rows