一個開放的配方：透過模型合併在一天內將特定語言的LLMs調整為推理模型

An Open Recipe: Adapting Language-Specific LLMs to a Reasoning Model in One Day via Model Merging

February 13, 2025

作者: Kunat Pipatanakul, Pittawat Taveekitworachai, Potsawee Manakul, Kasima Tharnpipitchai

cs.AI

摘要

本文探討資料選擇和模型合併方法，旨在將像 DeepSeek R1 這樣的先進推理能力納入特定語言的大型語言模型（LLMs），特別聚焦於泰語LLM。我們的目標是增強特定語言LLMs的推理能力，同時保持其目標語言能力。DeepSeek R1 在推理方面表現出色，但主要受益於高資源語言，如英語和中文。然而，由於以英語為中心的訓練數據和模型優化佔主導地位，低資源語言仍未得到應有的服務，這限制了這些語言的性能。這種限制導致代碼切換不可靠，並且在低資源語言的任務上效果不佳。與此同時，當地和區域LLM倡議已嘗試彌合這一差距，通過開發專注於提高當地語言忠實度的特定語言LLMs。我們證明，僅使用公開可用的數據集和120美元的計算預算，就可以增強特定語言LLMs的推理能力，使其與DeepSeek R1的水平相匹敵，同時不損害其在目標語言任務上的性能。

English

This paper investigates data selection and model merging methodologies aimed at incorporating advanced reasoning capabilities such as those of DeepSeek R1 into language-specific large language models (LLMs), with a particular focus on the Thai LLM. Our goal is to enhance the reasoning capabilities of language-specific LLMs while maintaining their target language abilities. DeepSeek R1 excels in reasoning but primarily benefits high-resource languages such as English and Chinese. However, low-resource languages remain underserved due to the dominance of English-centric training data and model optimizations, which limit performance in these languages. This limitation results in unreliable code-switching and diminished effectiveness on tasks in low-resource languages. Meanwhile, local and regional LLM initiatives have attempted to bridge this gap by developing language-specific LLMs that focus on improving local linguistic fidelity. We demonstrate that, with only publicly available datasets and a computational budget of $120, it is possible to enhance the reasoning capabilities of language-specific LLMs to match the level of DeepSeek R1, without compromising their performance on target language tasks.