WangchanThaiInstruct: An instruction-following Dataset for Culture-Aware, Multitask, and Multi-domain Evaluation in Thai
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915502198095872 |
|---|---|
| author | Limkonchotiwat, Peerat Tuchinda, Pume Lowphansirikul, Lalita Nonesung, Surapon Tasawong, Panuthep Aji, Alham Fikri Udomcharoenchaikit, Can Nutanong, Sarana |
| author_facet | Limkonchotiwat, Peerat Tuchinda, Pume Lowphansirikul, Lalita Nonesung, Surapon Tasawong, Panuthep Aji, Alham Fikri Udomcharoenchaikit, Can Nutanong, Sarana |
| contents | Large language models excel at instruction-following in English, but their performance in low-resource languages like Thai remains underexplored. Existing benchmarks often rely on translations, missing cultural and domain-specific nuances needed for real-world use. We present WangchanThaiInstruct, a human-authored Thai dataset for evaluation and instruction tuning, covering four professional domains and seven task types. Created through a multi-stage quality control process with annotators, domain experts, and AI researchers, WangchanThaiInstruct supports two studies: (1) a zero-shot evaluation showing performance gaps on culturally and professionally specific tasks, and (2) an instruction tuning study with ablations isolating the effect of native supervision. Models fine-tuned on WangchanThaiInstruct outperform those using translated data in both in-domain and out-of-domain benchmarks. These findings underscore the need for culturally and professionally grounded instruction data to improve LLM alignment in low-resource, linguistically diverse settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_15239 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | WangchanThaiInstruct: An instruction-following Dataset for Culture-Aware, Multitask, and Multi-domain Evaluation in Thai Limkonchotiwat, Peerat Tuchinda, Pume Lowphansirikul, Lalita Nonesung, Surapon Tasawong, Panuthep Aji, Alham Fikri Udomcharoenchaikit, Can Nutanong, Sarana Computation and Language Large language models excel at instruction-following in English, but their performance in low-resource languages like Thai remains underexplored. Existing benchmarks often rely on translations, missing cultural and domain-specific nuances needed for real-world use. We present WangchanThaiInstruct, a human-authored Thai dataset for evaluation and instruction tuning, covering four professional domains and seven task types. Created through a multi-stage quality control process with annotators, domain experts, and AI researchers, WangchanThaiInstruct supports two studies: (1) a zero-shot evaluation showing performance gaps on culturally and professionally specific tasks, and (2) an instruction tuning study with ablations isolating the effect of native supervision. Models fine-tuned on WangchanThaiInstruct outperform those using translated data in both in-domain and out-of-domain benchmarks. These findings underscore the need for culturally and professionally grounded instruction data to improve LLM alignment in low-resource, linguistically diverse settings. |
| title | WangchanThaiInstruct: An instruction-following Dataset for Culture-Aware, Multitask, and Multi-domain Evaluation in Thai |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2508.15239 |