FigLang2022 SharedTask on Understanding Figurative Language Leaderboard (v1.0)

Rank

Model

Acc@0

Acc@50

Acc@60

 

T5-3B (resource-rich baseline)

76.72 69.05 44.27

Submit to this leaderboard: You can submit your prediction by going to CodaLab.You will have to provide details about your underlying model and the number of parameters. All details on submission file format and a evaluation script will be provided there

We use Accuracy@60 for ranking all methods because it means the explanations are of superior quality.We will also conduct human evaluation of top 5 performing systems. When accounting for top 5 sytems we will take into account both systems with good Accuracy measures but also systems that are competitive but typically use smaller models (in terms of number of parameters)

Misc.

Citation