LLaMAX2：您的翻译增强模型在推理任务中同样表现出色

LLaMAX2: Your Translation-Enhanced Model also Performs Well in Reasoning

October 10, 2025

作者: Changjiang Gao, Zixian Huang, Jingyang Gong, Shujian Huang, Lei Li, Fei Yuan

cs.AI

摘要

通用大型语言模型（LLMs）在推理方面表现出色，但专为翻译优化的模型在处理推理任务时却显不足。为解决这一问题，我们提出了一种新颖的翻译增强方案，该方案始于指令模型，并仅在平行数据上实施层次选择性调优。遵循这一流程，我们推出了Qwen3-XPlus模型，该模型在高资源和低资源语言的翻译性能上均展现出显著提升，在如斯瓦希里语等低资源语言中实现了超过15的spBLEU和40以上的xComET得分。值得注意的是，仅使用小型平行数据集进行训练，Qwen3-XPlus在7项多语言任务上平均提升了1分以上，同时在15个主流推理数据集上保持了与Qwen3指令模型相当的水平。这项工作为多语言增强提供了一条前景广阔的路径，显著降低了复杂度，并提升了更广泛语言的可及性。代码与模型均已公开。

English

General Large Language Models (LLMs) excel in reasoning, but those enhanced for translation struggle with reasoning tasks. To address this, we propose a novel translationenhanced recipe that begins with instruct models and applies layer-selective tuning only on parallel data. Following this pipeline, we introduce the Qwen3-XPlus models, which demonstrate significant improvements in translation performance across both high- and lowresource languages, achieving 15+ spBLEU and 40+ xComet in low-resource languages, like Swahili. Interestingly, training only with small parallel datasets, Qwen3-XPlus achieves an average improvement of 1+ points on 7 multilingual tasks while maintaining proficiency comparable to the Qwen3 instruct model in 15 popular reasoning datasets. This work offers a promising approach to multilingual enhancement, significantly reducing complexity and enhancing accessibility for a wider range of languages. The code and model are publicly available.