List your company

Figurative Language Detection on IRFL

46.4Idioms AccuracyGRU
14.78422.99231.239.408Jun 3, 2026

Evaluation results

MethodLinks
GRUInput Modalities=Image vs Caption ⊕ Definition2026.0646.457.444.9249.56
gMLPInput Modalities=Image vs Caption ⊕ Definition2026.0643.864.5544.9251.36
RePercENTInput Modalities=Image vs Caption ⊕ Definition2026.0642.565.4944.9851.38
CLIPBackbone=ViT-B/32, Protocol=finetune-all, Input Modalities=Image vs Caption ⊕ Definition2026.0638.170.442.1650.81
GRU2026.0636.148.8133.6939.46
CLIPBackbone=ViT-B/32, Protocol=finetune-proj, Input Modalities=Image vs Caption ⊕ Definition2026.063666.541.4448.67
gMLP2026.0633.754.3728.2338.52
RePercENTSE (Semantic Encodings)=true, GSA (Group Slot Attention)=true2026.0633.661.8835.5644.07
RePercENTSE (Semantic Encodings)=true, GSA (Group Slot Attention)=false2026.0632.661.8831.9542.35
RePercENTSE (Semantic Encodings)=false, GSA (Group Slot Attention)=true2026.0630.263.1130.4541.56
RePercENTSE (Semantic Encodings)=false, GSA (Group Slot Attention)=false2026.0628.263.9727.8740.3
CLIPBackbone=ViT-B/32, Protocol=zero-shot, Input Modalities=Image vs Caption ⊕ Definition2026.062648.3836.0438.78
CLIP-ViT-B/32Evaluation Protocol=end-to-end FT2026.0622.570.1829.6141.73
CLIP-ViT-B/32Evaluation Protocol=projection-only FT2026.0620.468.3831.5941.41
CLIP-ViT-B/32Evaluation Protocol=zero-shot2026.061645.4923.4229.14