Multilinguality of Large Language Models From a Structural Perspective
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913178157318144 |
|---|---|
| author | Sakajo, Haruki Sakai, Yusuke Kamigaito, Hidetaka Watanabe, Taro |
| author_facet | Sakajo, Haruki Sakai, Yusuke Kamigaito, Hidetaka Watanabe, Taro |
| contents | Large language models (LLMs) have excelled in processing multiple languages through pre- and post-training on multilingual data, even though English dominates the training data. Prior work focusing on token representations has revealed how those LLMs process non-English text. Although these analyses have provided insightful findings, they fail to capture a structural view, which is an inherent property of language. In this study, we explore the multilinguality of LLMs through representational structural analysis. Our findings reveal that low-resource languages are structurally more different from English than high- and mid-resource languages, and that language-specific post-training alters their structures while preserving inter-language relationships. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2606_01800 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Multilinguality of Large Language Models From a Structural Perspective Sakajo, Haruki Sakai, Yusuke Kamigaito, Hidetaka Watanabe, Taro Computation and Language Artificial Intelligence Machine Learning Large language models (LLMs) have excelled in processing multiple languages through pre- and post-training on multilingual data, even though English dominates the training data. Prior work focusing on token representations has revealed how those LLMs process non-English text. Although these analyses have provided insightful findings, they fail to capture a structural view, which is an inherent property of language. In this study, we explore the multilinguality of LLMs through representational structural analysis. Our findings reveal that low-resource languages are structurally more different from English than high- and mid-resource languages, and that language-specific post-training alters their structures while preserving inter-language relationships. |
| title | Multilinguality of Large Language Models From a Structural Perspective |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2606.01800 |