List your company
ResearchFormerly WizWand

Commonsense Question Answering on CommonsenseQA (CSQA2) 2.0 (dev)

58.72AccuracyELABOR
51.845653.630355.41557.1997Sep 2, 2022
Updated 8mo ago

Evaluation results

MethodLinks
ELABORAnswer predictor=T5-large, Elaboration model=GPT2-large2022.0958.72
GPT-3Answer predictor=T5-large2022.0958.56
scratchAnswer predictor=T5-large, Elaboration model=GPT2-large2022.0956.99
pipelineAnswer predictor=T5-large, Elaboration model=GPT2-large2022.0956.63
selftalkAnswer predictor=T5-large2022.0955.88
vanillaAnswer predictor=T5-large2022.0955.25
WikipediaAnswer predictor=T5-large2022.0952.14
COMETAnswer predictor=T5-large2022.0952.11