Share your thoughts, 1 month free Claude Pro on usSee more

Analogy recognition on UNIT 2

88.32Accuracy

DeBERTa-v3 + ANALOGYKB

Updated 1mo ago

Evaluation Results

Method	Links
DeBERTa-v3 + ANALOGYKB 2023.05		88.32
Human 2023.05		87.5
RoBERTa-Large + ANALOGYKB 2023.05		87.28
Word Embedding from InstructGPT 2023.05		65.39
Word Embedding from RoBERTa-Large 2023.05		58.3
Sentence Embedding from SimCSE 2023.05		54.29
Sentence Embedding from SentenceBERT 2023.05		53.4
DeBERTa-v3 2023.05		50
RoBERTa-Large 2023.05		46.05
DeBERTa-v3 + ANALOGYKB (w/o check) 2023.05		45.18
RoBERTa-Large + ANALOGYKB (w/o check) 2023.05		44.2
T5-Large 2023.05		34.65
LUKE 2023.05		34.64
ERNIE 2023.05		34.21
BERT-Large 2023.05		32.89