Adapting Language-Specific LLMs to a Reasoning Model in One Day via Model Merging -- An Open Recipe
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917968924901376 |
|---|---|
| author | Pipatanakul, Kunat Taveekitworachai, Pittawat Manakul, Potsawee Tharnpipitchai, Kasima |
| author_facet | Pipatanakul, Kunat Taveekitworachai, Pittawat Manakul, Potsawee Tharnpipitchai, Kasima |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_09056 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Adapting Language-Specific LLMs to a Reasoning Model in One Day via Model Merging -- An Open Recipe Pipatanakul, Kunat Taveekitworachai, Pittawat Manakul, Potsawee Tharnpipitchai, Kasima Computation and Language Artificial Intelligence 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. |
| title | Adapting Language-Specific LLMs to a Reasoning Model in One Day via Model Merging -- An Open Recipe |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2502.09056 |