Skip to Main Content
Table 11: 
Accuracy (%) by question type on the annotated development subset.
Question TypeFTLM + +GBDT + +
Matching 57.0 68.1 
Reasoning 56.8 49.4 
 Summary 73.6 47.1 
 Logic 55.0 49.7 
 Arithmetic 30.2 24.5 
 Commonsense 53.4 41.7 
Single sentence 56.5 63.3 
Multiple sentences 56.9 49.5 
Question TypeFTLM + +GBDT + +
Matching 57.0 68.1 
Reasoning 56.8 49.4 
 Summary 73.6 47.1 
 Logic 55.0 49.7 
 Arithmetic 30.2 24.5 
 Commonsense 53.4 41.7 
Single sentence 56.5 63.3 
Multiple sentences 56.9 49.5 
Close Modal

or Create an Account

Close Modal
Close Modal