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

COPA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Common-sense ReasoningCOPA
Accuracy99.2
256
Question AnsweringCOPA
Accuracy96
59
Commonsense ReasoningCOPA (test)
Accuracy98.67
54
Causal ReasoningCOPA
Accuracy90
51
Sentence CompletionCOPA
Accuracy92.88
48
Multiple ChoiceCOPA
Accuracy100
36
Causal Question AnsweringCOPA
EM99.3
32
Multi-class ClassificationCopa
Accuracy92
22
Causal ReasoningCopa100
Accuracy83
12
Commonsense Causal ReasoningCOPA (dev)
Accuracy93
7
Commonsense reasoningBalanced COPA
Accuracy70.7
6
Commonsense ReasoningCOPA 2011
Accuracy79
6
Choice of Plausible AlternativesCOPA 11 languages
Score55.5
5
Finetuning domain recoveryCOPA
Recovery Score (Grader 1)5
4
Inference correction review (discard)COPA
MHA100
4
Natural Language InferenceCOPA
Accuracy80
3
Commonsense Causal ReasoningCOPA 5-shot
Accuracy85
3
Commonsense ReasoningCOPA es
Accuracy54.4
3
Commonsense ReasoningCOPA en
Accuracy56
3
Commonsense ReasoningCOPA (dev)
Accuracy86
3
Choice of Plausible AlternativesCOPA (dev)
Accuracy65.8
3
Inference correction review (correction)COPA
MHA100
2
Inference correction review (reason)COPA
MHA100
2
Timing comparisonCOPA
MHA45.5
2
Event/state classificationCOPA
MHA90.9
2
Showing 25 of 26 rows