Language Models

LEXTREME: A Multi-Lingual and Multi-Task Benchmark for the Legal Domain
LEXTREME gathers 11 legal datasets in 24 languages into one benchmark with a single score, and shows that legal language understanding remains hard.
Joel Niklaus, Veton Matoshi, Pooja Rani, Andrea Galassi, Matthias Stürmer, Ilias Chalkidis
Cite DOI URL
Fast Vocabulary Transfer for Language Model Compression
Fast Vocabulary Transfer moves a language model to an in-domain tokenizer, so the model runs faster and can shrink, with little loss in F1.
Leonidas Gee, Andrea Zugarini, Leonardo Rigutini, Paolo Torroni
Cite DOI URL