How Transliterations Improve Crosslingual Alignment

[Free GPT-4]
[DeepSeek]
[PDF] arxiv.org

MEXA: Multilingual Evaluation of English-Centric LLMs via Cross-Lingual Alignment

AH Kargaran, A Modarressi, N Nikeghbal… - arxiv preprint arxiv …, 2024 - arxiv.org
English-centric large language models (LLMs) often show strong multilingual capabilities.
However, the multilingual performance of these models remains unclear and is not …
Zapisz Cytuj Cytowane przez 2 Powiązane artykuły Wszystkie wersje 2 Wersja HTML