List your company
ResearchFormerly WizWand

Commonsense Question Answering on Scientific Commonsense (QASC) 1.0 (dev)

55.18AccuracyGPT-3
47.879249.774651.6753.5654Sep 2, 2022
Updated 8mo ago

Evaluation results

MethodLinks
GPT-3Answer predictor=T5-large2022.0955.18
ELABORAnswer predictor=T5-large, Elaboration model=GPT2-large2022.0954.21
pipelineAnswer predictor=T5-large, Elaboration model=GPT2-large2022.0952.48
scratchAnswer predictor=T5-large, Elaboration model=GPT2-large2022.0950.65
selftalkAnswer predictor=T5-large2022.0950.22
COMETAnswer predictor=T5-large2022.0949.35
vanillaAnswer predictor=T5-large2022.0948.49
WikipediaAnswer predictor=T5-large2022.0948.16