Aggregate Suite
Benchmarks
TaskDatasetSOTA resultTrendResults
Aggregate Suite
78.03Average Score
14
Aggregate Suite (test)
82.2Macro Accuracy
14
Aggregate Suite Metis splitting (Overall)
80.9Accuracy
11
Aggregate Suite PIQA, HellaSwag, WinoGrande, ARC-e, ARC-c
69Average Score
10