Publication:

LLM-based Translation for Latin: Summaries Improve Machine Translation

Date

Date

Date
2025
Conference or Workshop Item
Epub ahead of print
Simple data Full metadata Statistics
cris.virtual.orcidhttps://orcid.org/0000-0002-2063-4516
cris.virtualsource.orcid8fbbe5f4-ab2a-4bbe-a533-e6a4112e86d8
dc.date.accessioned2025-07-22T08:11:01Z
dc.date.available2025-07-22T08:11:01Z
dc.date.issued2025-05-15
dc.description.abstract

Recent studies demonstrated that modern Large Language Models set a new state-of-the-art in translating historical Latin texts into English and German. Building upon this foundation, we investigate the impact of incorporating text summaries into prompts for LLM-based translation tasks. Having both the historical text and a modern-language summary is a typical setup for classical editions. Our findings reveal that integrating summaries significantly enhances translation accuracy and coherence.

dc.identifier.urihttps://www.zora.uzh.ch/handle/20.500.14742/232136
dc.language.isoeng
dc.subjectLarge Language Models
dc.subjectMachine Translation
dc.subjectSummaries
dc.subjectLatin
dc.subjectHistorical Letters
dc.subject.ddc410 Linguistics
dc.subject.ddc000 Computer science, knowledge & systems
dc.subject.ddc400 Language
dc.title

LLM-based Translation for Latin: Summaries Improve Machine Translation

dc.typeconference_item
dcterms.accessRightsinfo:eu-repo/semantics/openAccess
dcterms.bibliographicCitation.originalpublishernameAssociation for Computational Linguistics
dspace.entity.typePublicationen
oairecerif.event.endDate2025-05-15
oairecerif.event.placeWinterthur
oairecerif.event.startDate2025-05-14
uzh.contributor.authorFischer, Dominic P
uzh.contributor.authorVolk, Martin
uzh.contributor.correspondenceYes
uzh.contributor.correspondenceNo
uzh.document.availabilitypostprint
uzh.eprint.datestamp2025-07-22 08:11:01
uzh.eprint.lastmod2025-07-22 08:21:23
uzh.eprint.statusChange2025-07-22 08:11:01
uzh.event.presentationTypeother
uzh.event.titleSwissText 2025
uzh.event.typeconference
uzh.harvester.ethYes
uzh.harvester.nbNo
uzh.identifier.doi10.5167/uzh-279356
uzh.oastatus.zoraGreen
uzh.publication.citationFischer, D. P., & Volk, M. (2025). LLM-based Translation for Latin: Summaries Improve Machine Translation. Proceedings of the Swiss Text Analytics Conference. Presented at the SwissText 2025, Association for Computational Linguistics.
uzh.publication.citationFischer, Dominic P; Volk, Martin (2025). LLM-based Translation for Latin: Summaries Improve Machine Translation. In: SwissText 2025, Winterthur, 14 Mai 2025 - 15 Mai 2025, Association for Computational Linguistics.
uzh.publication.citationFischer, D. P., & Volk, M. (2025). LLM-based Translation for Latin: Summaries Improve Machine Translation. Proceedings of the Swiss Text Analytics Conference. Presented at the SwissText 2025, Association for Computational Linguistics.
uzh.publication.freeAccessAtUNSPECIFIED
uzh.publication.originalworkoriginal
uzh.publication.publishedStatusfirstelectronic
uzh.publication.seriesTitleProceedings of the Swiss Text Analytics Conference
uzh.workflow.eprintid279356
uzh.workflow.fulltextStatuspublic
uzh.workflow.revisions16
uzh.workflow.rightsCheckkeininfo
uzh.workflow.statusarchive
Files

Original bundle

Name:
LLM_based_MT__Summaries_Improve_Translation_Quality.pdf
Size:
111.19 KB
Format:
Adobe Portable Document Format
LLM_based_MT__Summaries_Improve_Translation_Quality.pdfview fileDownload |Download111.19 KB
Publication available in collections:
Publications of Institute of Computational Linguistics
Publications of Zurich Center for Linguistics