List your company

Cross-language Vocabulary Overlap on 6 languages All transfers

0.74JSDUnigram
0.72960.79980.870.9402May 26, 2023

Evaluation results

MethodLinks
UnigramTokenizer=Unigram2023.050.74
TokMixTokenizer=TokMix2023.050.77
BPETokenizer=BPE2023.050.8
NoOverlapTokenizer=NoOverlap2023.051