LaMP-4
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
LaMP-4 v1 (test)
17.8ROUGE-1
22
LaMP-4
21.9ROUGE-1
15
LaMP-4 (val)
19.3ROUGE-1 Recall
14
LaMP-4 time-based split (dev)
21ROUGE-1
8
LAMP-4 OOD
21ROUGE-1
7
LaMP-4 (random)
16.7ROUGE-1
7
LaMP-4 (test)
19R-1 Score
7
LaMP-4 (test)
40Alignment Win (%)
2