作者位次 #4/5ACL 2026Findings· Findings of ACL
Data-Centric Continual Pre-training for 500+ Languages: A New Bilingual Translation Corpus and Multilingual Models
Shaoxiong Ji · Zihao Li · Jaakko Paavola · Hengyu Luo · Jörg Tiedemann
摘要
This paper investigates a critical design decision in the practice of massively multilingual continual pre-training — the inclusion of parallel data. Specifically, we study the impact of bilingual translation data for massively multilingual language adaptation of the Llama 3 family of models to 5...