Beyond Decoder-only: Large Language Models Can be Good Encoders for Machine Translation Paper • 2503.06594 • Published Mar 9 • 5
XALMA-13B-Pretrain + Separate Training Collection Fifty expert models are produced by separately fine-tuning XALMA-13B-Pretrain on each of 50 languages. • 50 items • Updated Feb 23 • 1
Asymmetric Conflict and Synergy in Post-training for LLM-based Multilingual Machine Translation Paper • 2502.11223 • Published Feb 16 • 1