Skip to Main Content
Table 9: 

QA accuracy (with standard deviation across 7 experiments), and retrieval performance, measured by Recall@10, of baseline models on the test set.

ModelAccuracyRecall@10
Majority 53.9 
RoBERTa* 63.6 ± 1.3 
RoBERTaIR-Q 53.6 ± 1.0 0.174 
RoBERTa*IR-Q 63.6 ± 1.0 0.174 
RoBERTa*IR-D 61.7 ± 2.2 0.195 
RoBERTa*IR-ORA-D 62.0 ± 1.3 0.282 
RoBERTa*ORA-P 70.7 ± 0.6 
RoBERTa*ORA-P-Dlast-step-raw 65.2 ± 1.4 
RoBERTa*ORA-P-Dlast-step 72.0 ± 1.0 
ModelAccuracyRecall@10
Majority 53.9 
RoBERTa* 63.6 ± 1.3 
RoBERTaIR-Q 53.6 ± 1.0 0.174 
RoBERTa*IR-Q 63.6 ± 1.0 0.174 
RoBERTa*IR-D 61.7 ± 2.2 0.195 
RoBERTa*IR-ORA-D 62.0 ± 1.3 0.282 
RoBERTa*ORA-P 70.7 ± 0.6 
RoBERTa*ORA-P-Dlast-step-raw 65.2 ± 1.4 
RoBERTa*ORA-P-Dlast-step 72.0 ± 1.0 
Close Modal

or Create an Account

Close Modal
Close Modal