1
Speculative Decoding Across Languages
arXiv最新论文提出跨语言推测解码方法，无需目标语言训练数据即可加速多语言大模型推理
arXiv:2605.30580v1 Announce Type: cross Abstract: Speculative decoding has become a crucial component of large language model (LLM) inference, enablin…
2
Polite on the Surface, Wrong in Practice: A Curated Dataset for Fixing Honorific Failures in Multilingual Bangla Generation
孟加拉语生成中敬语错误屡见不鲜，这份精心策划的数据集专门针对多语言模型进行修复，值得关注。
arXiv:2605.22487v1 Announce Type: new Abstract: Recent advances in Multilingual Large Language Models (MLLMs) have significantly enhanced cross-lingua…