List your company

Math Word Problem solving on UnbiasedMWP 1:N Chinese (test)

48.4AccuracyATHENA
13.5622.60531.6540.695Nov 2, 2023

Evaluation results

MethodLinks
ATHENABackbone=RoBERTa-large2023.1148.4
ATHENABackbone=RoBERTa-base2023.1135.4
DeductReasonerBackbone=RoBERTa-large2023.1133.7
DeductReasonerBackbone=RoBERTa-base2023.1126.5
Graph-to-TreeBackbone=Random embedding2023.1124.3
GTSBackbone=Random embedding2023.1122.8
TransformerBackbone=Random embedding2023.1116.9
R-TransformerBackbone=RoBERTa-base2023.1114.9