The Effect of Language Diversity When Fine-Tuning Large Language Models for Translation Article Swipe
YOU?
·
· 2025
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2505.13090
Prior research diverges on language diversity in LLM fine-tuning: Some studies report benefits while others find no advantages. Through controlled fine-tuning experiments across 132 translation directions, we systematically resolve these disparities. We find that expanding language diversity during fine-tuning improves translation quality for both unsupervised and -- surprisingly -- supervised pairs, despite less diverse models being fine-tuned exclusively on these supervised pairs. However, benefits plateau or decrease beyond a certain diversity threshold. We show that increased language diversity creates more language-agnostic representations. These representational adaptations help explain the improved performance in models fine-tuned with greater diversity.
Related Topics
- Type
- preprint
- Language
- en
- Landing Page
- http://arxiv.org/abs/2505.13090
- https://arxiv.org/pdf/2505.13090
- OA Status
- green
- OpenAlex ID
- https://openalex.org/W4417287168
Raw OpenAlex JSON
- OpenAlex ID
-
https://openalex.org/W4417287168Canonical identifier for this work in OpenAlex
- DOI
-
https://doi.org/10.48550/arxiv.2505.13090Digital Object Identifier
- Title
-
The Effect of Language Diversity When Fine-Tuning Large Language Models for TranslationWork title
- Type
-
preprintOpenAlex work type
- Language
-
enPrimary language
- Publication year
-
2025Year of publication
- Publication date
-
2025-05-19Full publication date if available
- Authors
-
David Stap, Christof MonzList of authors in order
- Landing page
-
https://arxiv.org/abs/2505.13090Publisher landing page
- PDF URL
-
https://arxiv.org/pdf/2505.13090Direct link to full text PDF
- Open access
-
YesWhether a free full text is available
- OA status
-
greenOpen access status per OpenAlex
- OA URL
-
https://arxiv.org/pdf/2505.13090Direct OA link when available
- Cited by
-
0Total citation count in OpenAlex
Full payload
| id | https://openalex.org/W4417287168 |
|---|---|
| doi | https://doi.org/10.48550/arxiv.2505.13090 |
| ids.doi | https://doi.org/10.48550/arxiv.2505.13090 |
| ids.openalex | https://openalex.org/W4417287168 |
| fwci | |
| type | preprint |
| title | The Effect of Language Diversity When Fine-Tuning Large Language Models for Translation |
| biblio.issue | |
| biblio.volume | |
| biblio.last_page | |
| biblio.first_page | |
| is_xpac | False |
| apc_list | |
| apc_paid | |
| language | en |
| locations[0].id | pmh:oai:arXiv.org:2505.13090 |
| locations[0].is_oa | True |
| locations[0].source.id | https://openalex.org/S4306400194 |
| locations[0].source.issn | |
| locations[0].source.type | repository |
| locations[0].source.is_oa | True |
| locations[0].source.issn_l | |
| locations[0].source.is_core | False |
| locations[0].source.is_in_doaj | False |
| locations[0].source.display_name | arXiv (Cornell University) |
| locations[0].source.host_organization | https://openalex.org/I205783295 |
| locations[0].source.host_organization_name | Cornell University |
| locations[0].source.host_organization_lineage | https://openalex.org/I205783295 |
| locations[0].license | cc-by |
| locations[0].pdf_url | https://arxiv.org/pdf/2505.13090 |
| locations[0].version | submittedVersion |
| locations[0].raw_type | text |
| locations[0].license_id | https://openalex.org/licenses/cc-by |
| locations[0].is_accepted | False |
| locations[0].is_published | False |
| locations[0].raw_source_name | |
| locations[0].landing_page_url | http://arxiv.org/abs/2505.13090 |
| locations[1].id | doi:10.48550/arxiv.2505.13090 |
| locations[1].is_oa | True |
| locations[1].source.id | https://openalex.org/S4306400194 |
| locations[1].source.issn | |
| locations[1].source.type | repository |
| locations[1].source.is_oa | True |
| locations[1].source.issn_l | |
| locations[1].source.is_core | False |
| locations[1].source.is_in_doaj | False |
| locations[1].source.display_name | arXiv (Cornell University) |
| locations[1].source.host_organization | https://openalex.org/I205783295 |
| locations[1].source.host_organization_name | Cornell University |
| locations[1].source.host_organization_lineage | https://openalex.org/I205783295 |
| locations[1].license | cc-by |
| locations[1].pdf_url | |
| locations[1].version | |
| locations[1].raw_type | article |
| locations[1].license_id | https://openalex.org/licenses/cc-by |
| locations[1].is_accepted | False |
| locations[1].is_published | |
| locations[1].raw_source_name | |
| locations[1].landing_page_url | https://doi.org/10.48550/arxiv.2505.13090 |
| indexed_in | arxiv, datacite |
| authorships[0].author.id | https://openalex.org/A5015155057 |
| authorships[0].author.orcid | |
| authorships[0].author.display_name | David Stap |
| authorships[0].author_position | first |
| authorships[0].raw_author_name | Stap, David |
| authorships[0].is_corresponding | False |
| authorships[1].author.id | https://openalex.org/A5109059955 |
| authorships[1].author.orcid | |
| authorships[1].author.display_name | Christof Monz |
| authorships[1].author_position | last |
| authorships[1].raw_author_name | Monz, Christof |
| authorships[1].is_corresponding | False |
| has_content.pdf | True |
| has_content.grobid_xml | True |
| is_paratext | False |
| open_access.is_oa | True |
| open_access.oa_url | https://arxiv.org/pdf/2505.13090 |
| open_access.oa_status | green |
| open_access.any_repository_has_fulltext | False |
| created_date | 2025-10-10T00:00:00 |
| display_name | The Effect of Language Diversity When Fine-Tuning Large Language Models for Translation |
| has_fulltext | True |
| is_retracted | False |
| updated_date | 2025-12-13T09:19:10.953794 |
| primary_topic | |
| cited_by_count | 0 |
| locations_count | 2 |
| best_oa_location.id | pmh:oai:arXiv.org:2505.13090 |
| best_oa_location.is_oa | True |
| best_oa_location.source.id | https://openalex.org/S4306400194 |
| best_oa_location.source.issn | |
| best_oa_location.source.type | repository |
| best_oa_location.source.is_oa | True |
| best_oa_location.source.issn_l | |
| best_oa_location.source.is_core | False |
| best_oa_location.source.is_in_doaj | False |
| best_oa_location.source.display_name | arXiv (Cornell University) |
| best_oa_location.source.host_organization | https://openalex.org/I205783295 |
| best_oa_location.source.host_organization_name | Cornell University |
| best_oa_location.source.host_organization_lineage | https://openalex.org/I205783295 |
| best_oa_location.license | cc-by |
| best_oa_location.pdf_url | https://arxiv.org/pdf/2505.13090 |
| best_oa_location.version | submittedVersion |
| best_oa_location.raw_type | text |
| best_oa_location.license_id | https://openalex.org/licenses/cc-by |
| best_oa_location.is_accepted | False |
| best_oa_location.is_published | False |
| best_oa_location.raw_source_name | |
| best_oa_location.landing_page_url | http://arxiv.org/abs/2505.13090 |
| primary_location.id | pmh:oai:arXiv.org:2505.13090 |
| primary_location.is_oa | True |
| primary_location.source.id | https://openalex.org/S4306400194 |
| primary_location.source.issn | |
| primary_location.source.type | repository |
| primary_location.source.is_oa | True |
| primary_location.source.issn_l | |
| primary_location.source.is_core | False |
| primary_location.source.is_in_doaj | False |
| primary_location.source.display_name | arXiv (Cornell University) |
| primary_location.source.host_organization | https://openalex.org/I205783295 |
| primary_location.source.host_organization_name | Cornell University |
| primary_location.source.host_organization_lineage | https://openalex.org/I205783295 |
| primary_location.license | cc-by |
| primary_location.pdf_url | https://arxiv.org/pdf/2505.13090 |
| primary_location.version | submittedVersion |
| primary_location.raw_type | text |
| primary_location.license_id | https://openalex.org/licenses/cc-by |
| primary_location.is_accepted | False |
| primary_location.is_published | False |
| primary_location.raw_source_name | |
| primary_location.landing_page_url | http://arxiv.org/abs/2505.13090 |
| publication_date | 2025-05-19 |
| publication_year | 2025 |
| referenced_works_count | 0 |
| abstract_inverted_index.a | 68 |
| abstract_inverted_index.-- | 46, 48 |
| abstract_inverted_index.We | 31, 72 |
| abstract_inverted_index.in | 6, 90 |
| abstract_inverted_index.no | 16 |
| abstract_inverted_index.on | 3, 58 |
| abstract_inverted_index.or | 65 |
| abstract_inverted_index.we | 26 |
| abstract_inverted_index.132 | 23 |
| abstract_inverted_index.LLM | 7 |
| abstract_inverted_index.and | 45 |
| abstract_inverted_index.for | 42 |
| abstract_inverted_index.the | 87 |
| abstract_inverted_index.Some | 9 |
| abstract_inverted_index.both | 43 |
| abstract_inverted_index.find | 15, 32 |
| abstract_inverted_index.help | 85 |
| abstract_inverted_index.less | 52 |
| abstract_inverted_index.more | 79 |
| abstract_inverted_index.show | 73 |
| abstract_inverted_index.that | 33, 74 |
| abstract_inverted_index.with | 93 |
| abstract_inverted_index.Prior | 0 |
| abstract_inverted_index.These | 82 |
| abstract_inverted_index.being | 55 |
| abstract_inverted_index.these | 29, 59 |
| abstract_inverted_index.while | 13 |
| abstract_inverted_index.across | 22 |
| abstract_inverted_index.beyond | 67 |
| abstract_inverted_index.during | 37 |
| abstract_inverted_index.models | 54, 91 |
| abstract_inverted_index.others | 14 |
| abstract_inverted_index.pairs, | 50 |
| abstract_inverted_index.pairs. | 61 |
| abstract_inverted_index.report | 11 |
| abstract_inverted_index.Through | 18 |
| abstract_inverted_index.certain | 69 |
| abstract_inverted_index.creates | 78 |
| abstract_inverted_index.despite | 51 |
| abstract_inverted_index.diverse | 53 |
| abstract_inverted_index.explain | 86 |
| abstract_inverted_index.greater | 94 |
| abstract_inverted_index.plateau | 64 |
| abstract_inverted_index.quality | 41 |
| abstract_inverted_index.resolve | 28 |
| abstract_inverted_index.studies | 10 |
| abstract_inverted_index.However, | 62 |
| abstract_inverted_index.benefits | 12, 63 |
| abstract_inverted_index.decrease | 66 |
| abstract_inverted_index.diverges | 2 |
| abstract_inverted_index.improved | 88 |
| abstract_inverted_index.improves | 39 |
| abstract_inverted_index.language | 4, 35, 76 |
| abstract_inverted_index.research | 1 |
| abstract_inverted_index.diversity | 5, 36, 70, 77 |
| abstract_inverted_index.expanding | 34 |
| abstract_inverted_index.increased | 75 |
| abstract_inverted_index.controlled | 19 |
| abstract_inverted_index.diversity. | 95 |
| abstract_inverted_index.fine-tuned | 56, 92 |
| abstract_inverted_index.supervised | 49, 60 |
| abstract_inverted_index.threshold. | 71 |
| abstract_inverted_index.adaptations | 84 |
| abstract_inverted_index.advantages. | 17 |
| abstract_inverted_index.directions, | 25 |
| abstract_inverted_index.exclusively | 57 |
| abstract_inverted_index.experiments | 21 |
| abstract_inverted_index.fine-tuning | 20, 38 |
| abstract_inverted_index.performance | 89 |
| abstract_inverted_index.translation | 24, 40 |
| abstract_inverted_index.disparities. | 30 |
| abstract_inverted_index.fine-tuning: | 8 |
| abstract_inverted_index.surprisingly | 47 |
| abstract_inverted_index.unsupervised | 44 |
| abstract_inverted_index.systematically | 27 |
| abstract_inverted_index.representational | 83 |
| abstract_inverted_index.representations. | 81 |
| abstract_inverted_index.language-agnostic | 80 |
| cited_by_percentile_year | |
| countries_distinct_count | 0 |
| institutions_distinct_count | 2 |
| citation_normalized_percentile |