Few-Shot Multilingual Coreference Resolution Using Long-Context Large Language Models Article Swipe
Association for Computational Linguistics 2025
·
YOU?
·
· 2025
· Open Access
·
· DOI: https://doi.org/10.48448/r8e2-7k92
YOU?
·
· 2025
· Open Access
·
· DOI: https://doi.org/10.48448/r8e2-7k92
In this work, we present our system, which ranked second in the CRAC 2025 Shared Task on Multilingual Coreference Resolution (LLM Track). For multilingual coreference resolution, our system mainly uses long-context large language models (LLMs) in a few-shot in-context learning setting. Among the various approaches we explored, few-shot prompting proved to be the most effective, particularly due to the complexity of the task and the availability of high-quality data with referential relationships provided as part of the competition. We employed Gemini 2.5 Pro, one of the best available closed-source long-context LLMs at the time of submission. Our system achieved a CoNLL F1 score of 61.74 on the mini-testset, demonstrating that performance improves significantly with the number of few-shot examples provided, thanks to the model's extended context window. While this approach comes with trade-offs in terms of inference cost and response latency, it highlights the potential of long-context LLMs for tackling multilingual coreference without task-specific fine-tuning. Although direct comparisons with traditional supervised systems are not straightforward, our findings provide valuable insights and open avenues for future work, particularly in expanding support for low-resource languages.
Related Topics
Concepts
Metadata
- Type
- other
- Landing Page
- https://doi.org/10.48448/r8e2-7k92
- OA Status
- green
- OpenAlex ID
- https://openalex.org/W7106805704
All OpenAlex metadata
Raw OpenAlex JSON
- OpenAlex ID
-
https://openalex.org/W7106805704Canonical identifier for this work in OpenAlex
- DOI
-
https://doi.org/10.48448/r8e2-7k92Digital Object Identifier
- Title
-
Few-Shot Multilingual Coreference Resolution Using Long-Context Large Language ModelsWork title
- Type
-
otherOpenAlex work type
- Publication year
-
2025Year of publication
- Publication date
-
2025-10-25Full publication date if available
- Authors
-
Association for Computational Linguistics 2025List of authors in order
- Landing page
-
https://doi.org/10.48448/r8e2-7k92Publisher landing page
- Open access
-
YesWhether a free full text is available
- OA status
-
greenOpen access status per OpenAlex
- OA URL
-
https://doi.org/10.48448/r8e2-7k92Direct OA link when available
- Concepts
-
Coreference, Computer science, Task (project management), Natural language processing, Artificial intelligence, Inference, Context (archaeology), Resolution (logic), Training set, Language model, Computational linguistics, Language understanding, Machine learning, Labeled data, Deep learning, Task analysisTop concepts (fields/topics) attached by OpenAlex
- Cited by
-
0Total citation count in OpenAlex
Full payload
| id | https://openalex.org/W7106805704 |
|---|---|
| doi | https://doi.org/10.48448/r8e2-7k92 |
| ids.doi | https://doi.org/10.48448/r8e2-7k92 |
| ids.openalex | https://openalex.org/W7106805704 |
| fwci | |
| type | other |
| title | Few-Shot Multilingual Coreference Resolution Using Long-Context Large Language Models |
| biblio.issue | |
| biblio.volume | |
| biblio.last_page | |
| biblio.first_page | |
| is_xpac | False |
| apc_list | |
| apc_paid | |
| concepts[0].id | https://openalex.org/C28076734 |
| concepts[0].level | 3 |
| concepts[0].score | 0.9764889478683472 |
| concepts[0].wikidata | https://www.wikidata.org/wiki/Q63087 |
| concepts[0].display_name | Coreference |
| concepts[1].id | https://openalex.org/C41008148 |
| concepts[1].level | 0 |
| concepts[1].score | 0.8245929479598999 |
| concepts[1].wikidata | https://www.wikidata.org/wiki/Q21198 |
| concepts[1].display_name | Computer science |
| concepts[2].id | https://openalex.org/C2780451532 |
| concepts[2].level | 2 |
| concepts[2].score | 0.705143392086029 |
| concepts[2].wikidata | https://www.wikidata.org/wiki/Q759676 |
| concepts[2].display_name | Task (project management) |
| concepts[3].id | https://openalex.org/C204321447 |
| concepts[3].level | 1 |
| concepts[3].score | 0.6862990260124207 |
| concepts[3].wikidata | https://www.wikidata.org/wiki/Q30642 |
| concepts[3].display_name | Natural language processing |
| concepts[4].id | https://openalex.org/C154945302 |
| concepts[4].level | 1 |
| concepts[4].score | 0.6518021821975708 |
| concepts[4].wikidata | https://www.wikidata.org/wiki/Q11660 |
| concepts[4].display_name | Artificial intelligence |
| concepts[5].id | https://openalex.org/C2776214188 |
| concepts[5].level | 2 |
| concepts[5].score | 0.6451393961906433 |
| concepts[5].wikidata | https://www.wikidata.org/wiki/Q408386 |
| concepts[5].display_name | Inference |
| concepts[6].id | https://openalex.org/C2779343474 |
| concepts[6].level | 2 |
| concepts[6].score | 0.6391147375106812 |
| concepts[6].wikidata | https://www.wikidata.org/wiki/Q3109175 |
| concepts[6].display_name | Context (archaeology) |
| concepts[7].id | https://openalex.org/C138268822 |
| concepts[7].level | 2 |
| concepts[7].score | 0.6159297227859497 |
| concepts[7].wikidata | https://www.wikidata.org/wiki/Q1051925 |
| concepts[7].display_name | Resolution (logic) |
| concepts[8].id | https://openalex.org/C51632099 |
| concepts[8].level | 2 |
| concepts[8].score | 0.4476976692676544 |
| concepts[8].wikidata | https://www.wikidata.org/wiki/Q3985153 |
| concepts[8].display_name | Training set |
| concepts[9].id | https://openalex.org/C137293760 |
| concepts[9].level | 2 |
| concepts[9].score | 0.42301681637763977 |
| concepts[9].wikidata | https://www.wikidata.org/wiki/Q3621696 |
| concepts[9].display_name | Language model |
| concepts[10].id | https://openalex.org/C155092808 |
| concepts[10].level | 2 |
| concepts[10].score | 0.3749261498451233 |
| concepts[10].wikidata | https://www.wikidata.org/wiki/Q182557 |
| concepts[10].display_name | Computational linguistics |
| concepts[11].id | https://openalex.org/C2983448237 |
| concepts[11].level | 2 |
| concepts[11].score | 0.3596290349960327 |
| concepts[11].wikidata | https://www.wikidata.org/wiki/Q1078276 |
| concepts[11].display_name | Language understanding |
| concepts[12].id | https://openalex.org/C119857082 |
| concepts[12].level | 1 |
| concepts[12].score | 0.3514323830604553 |
| concepts[12].wikidata | https://www.wikidata.org/wiki/Q2539 |
| concepts[12].display_name | Machine learning |
| concepts[13].id | https://openalex.org/C2776145971 |
| concepts[13].level | 2 |
| concepts[13].score | 0.3408384323120117 |
| concepts[13].wikidata | https://www.wikidata.org/wiki/Q30673951 |
| concepts[13].display_name | Labeled data |
| concepts[14].id | https://openalex.org/C108583219 |
| concepts[14].level | 2 |
| concepts[14].score | 0.2580905258655548 |
| concepts[14].wikidata | https://www.wikidata.org/wiki/Q197536 |
| concepts[14].display_name | Deep learning |
| concepts[15].id | https://openalex.org/C175154964 |
| concepts[15].level | 3 |
| concepts[15].score | 0.25446388125419617 |
| concepts[15].wikidata | https://www.wikidata.org/wiki/Q380077 |
| concepts[15].display_name | Task analysis |
| keywords[0].id | https://openalex.org/keywords/coreference |
| keywords[0].score | 0.9764889478683472 |
| keywords[0].display_name | Coreference |
| keywords[1].id | https://openalex.org/keywords/task |
| keywords[1].score | 0.705143392086029 |
| keywords[1].display_name | Task (project management) |
| keywords[2].id | https://openalex.org/keywords/inference |
| keywords[2].score | 0.6451393961906433 |
| keywords[2].display_name | Inference |
| keywords[3].id | https://openalex.org/keywords/context |
| keywords[3].score | 0.6391147375106812 |
| keywords[3].display_name | Context (archaeology) |
| keywords[4].id | https://openalex.org/keywords/resolution |
| keywords[4].score | 0.6159297227859497 |
| keywords[4].display_name | Resolution (logic) |
| keywords[5].id | https://openalex.org/keywords/training-set |
| keywords[5].score | 0.4476976692676544 |
| keywords[5].display_name | Training set |
| language | |
| locations[0].id | doi:10.48448/r8e2-7k92 |
| locations[0].is_oa | True |
| locations[0].source.id | https://openalex.org/S7407053148 |
| locations[0].source.type | repository |
| locations[0].source.is_oa | False |
| locations[0].source.issn_l | |
| locations[0].source.is_core | False |
| locations[0].source.is_in_doaj | False |
| locations[0].source.display_name | Underline Science Inc. |
| locations[0].source.host_organization | |
| locations[0].source.host_organization_name | |
| locations[0].license | |
| locations[0].pdf_url | |
| locations[0].version | |
| locations[0].raw_type | article |
| locations[0].license_id | |
| locations[0].is_accepted | False |
| locations[0].is_published | |
| locations[0].raw_source_name | |
| locations[0].landing_page_url | https://doi.org/10.48448/r8e2-7k92 |
| indexed_in | datacite |
| authorships[0].author.id | |
| authorships[0].author.orcid | |
| authorships[0].author.display_name | Association for Computational Linguistics 2025 |
| authorships[0].author_position | first |
| authorships[0].raw_author_name | Association for Computational Linguistics 2025 |
| authorships[0].is_corresponding | True |
| has_content.pdf | False |
| has_content.grobid_xml | False |
| is_paratext | False |
| open_access.is_oa | True |
| open_access.oa_url | https://doi.org/10.48448/r8e2-7k92 |
| open_access.oa_status | green |
| open_access.any_repository_has_fulltext | False |
| created_date | 2025-11-28T00:00:00 |
| display_name | Few-Shot Multilingual Coreference Resolution Using Long-Context Large Language Models |
| has_fulltext | False |
| is_retracted | False |
| updated_date | 2025-11-28T02:12:24.556248 |
| primary_topic | |
| cited_by_count | 0 |
| locations_count | 1 |
| best_oa_location.id | doi:10.48448/r8e2-7k92 |
| best_oa_location.is_oa | True |
| best_oa_location.source.id | https://openalex.org/S7407053148 |
| best_oa_location.source.type | repository |
| best_oa_location.source.is_oa | False |
| 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 | Underline Science Inc. |
| best_oa_location.source.host_organization | |
| best_oa_location.source.host_organization_name | |
| best_oa_location.license | |
| best_oa_location.pdf_url | |
| best_oa_location.version | |
| best_oa_location.raw_type | article |
| best_oa_location.license_id | |
| best_oa_location.is_accepted | False |
| best_oa_location.is_published | False |
| best_oa_location.raw_source_name | |
| best_oa_location.landing_page_url | https://doi.org/10.48448/r8e2-7k92 |
| primary_location.id | doi:10.48448/r8e2-7k92 |
| primary_location.is_oa | True |
| primary_location.source.id | https://openalex.org/S7407053148 |
| primary_location.source.type | repository |
| primary_location.source.is_oa | False |
| primary_location.source.issn_l | |
| primary_location.source.is_core | False |
| primary_location.source.is_in_doaj | False |
| primary_location.source.display_name | Underline Science Inc. |
| primary_location.source.host_organization | |
| primary_location.source.host_organization_name | |
| primary_location.license | |
| primary_location.pdf_url | |
| primary_location.version | |
| primary_location.raw_type | article |
| primary_location.license_id | |
| primary_location.is_accepted | False |
| primary_location.is_published | False |
| primary_location.raw_source_name | |
| primary_location.landing_page_url | https://doi.org/10.48448/r8e2-7k92 |
| publication_date | 2025-10-25 |
| publication_year | 2025 |
| referenced_works_count | 0 |
| abstract_inverted_index.a | 36, 99 |
| abstract_inverted_index.F1 | 101 |
| abstract_inverted_index.In | 0 |
| abstract_inverted_index.We | 78 |
| abstract_inverted_index.as | 73 |
| abstract_inverted_index.at | 91 |
| abstract_inverted_index.be | 51 |
| abstract_inverted_index.in | 10, 35, 133, 177 |
| abstract_inverted_index.it | 141 |
| abstract_inverted_index.of | 60, 66, 75, 84, 94, 103, 116, 135, 145 |
| abstract_inverted_index.on | 16, 105 |
| abstract_inverted_index.to | 50, 57, 121 |
| abstract_inverted_index.we | 3, 45 |
| abstract_inverted_index.2.5 | 81 |
| abstract_inverted_index.For | 22 |
| abstract_inverted_index.Our | 96 |
| abstract_inverted_index.and | 63, 138, 170 |
| abstract_inverted_index.are | 162 |
| abstract_inverted_index.due | 56 |
| abstract_inverted_index.for | 148, 173, 180 |
| abstract_inverted_index.not | 163 |
| abstract_inverted_index.one | 83 |
| abstract_inverted_index.our | 5, 26, 165 |
| abstract_inverted_index.the | 11, 42, 52, 58, 61, 64, 76, 85, 92, 106, 114, 122, 143 |
| abstract_inverted_index.(LLM | 20 |
| abstract_inverted_index.2025 | 13 |
| abstract_inverted_index.CRAC | 12 |
| abstract_inverted_index.LLMs | 90, 147 |
| abstract_inverted_index.Pro, | 82 |
| abstract_inverted_index.Task | 15 |
| abstract_inverted_index.best | 86 |
| abstract_inverted_index.cost | 137 |
| abstract_inverted_index.data | 68 |
| abstract_inverted_index.most | 53 |
| abstract_inverted_index.open | 171 |
| abstract_inverted_index.part | 74 |
| abstract_inverted_index.task | 62 |
| abstract_inverted_index.that | 109 |
| abstract_inverted_index.this | 1, 128 |
| abstract_inverted_index.time | 93 |
| abstract_inverted_index.uses | 29 |
| abstract_inverted_index.with | 69, 113, 131, 158 |
| abstract_inverted_index.61.74 | 104 |
| abstract_inverted_index.Among | 41 |
| abstract_inverted_index.CoNLL | 100 |
| abstract_inverted_index.While | 127 |
| abstract_inverted_index.comes | 130 |
| abstract_inverted_index.large | 31 |
| abstract_inverted_index.score | 102 |
| abstract_inverted_index.terms | 134 |
| abstract_inverted_index.which | 7 |
| abstract_inverted_index.work, | 2, 175 |
| abstract_inverted_index.(LLMs) | 34 |
| abstract_inverted_index.Gemini | 80 |
| abstract_inverted_index.Shared | 14 |
| abstract_inverted_index.direct | 156 |
| abstract_inverted_index.future | 174 |
| abstract_inverted_index.mainly | 28 |
| abstract_inverted_index.models | 33 |
| abstract_inverted_index.number | 115 |
| abstract_inverted_index.proved | 49 |
| abstract_inverted_index.ranked | 8 |
| abstract_inverted_index.second | 9 |
| abstract_inverted_index.system | 27, 97 |
| abstract_inverted_index.thanks | 120 |
| abstract_inverted_index.Track). | 21 |
| abstract_inverted_index.avenues | 172 |
| abstract_inverted_index.context | 125 |
| abstract_inverted_index.model's | 123 |
| abstract_inverted_index.present | 4 |
| abstract_inverted_index.provide | 167 |
| abstract_inverted_index.support | 179 |
| abstract_inverted_index.system, | 6 |
| abstract_inverted_index.systems | 161 |
| abstract_inverted_index.various | 43 |
| abstract_inverted_index.window. | 126 |
| abstract_inverted_index.without | 152 |
| abstract_inverted_index.Although | 155 |
| abstract_inverted_index.achieved | 98 |
| abstract_inverted_index.approach | 129 |
| abstract_inverted_index.employed | 79 |
| abstract_inverted_index.examples | 118 |
| abstract_inverted_index.extended | 124 |
| abstract_inverted_index.few-shot | 37, 47, 117 |
| abstract_inverted_index.findings | 166 |
| abstract_inverted_index.improves | 111 |
| abstract_inverted_index.insights | 169 |
| abstract_inverted_index.language | 32 |
| abstract_inverted_index.latency, | 140 |
| abstract_inverted_index.learning | 39 |
| abstract_inverted_index.provided | 72 |
| abstract_inverted_index.response | 139 |
| abstract_inverted_index.setting. | 40 |
| abstract_inverted_index.tackling | 149 |
| abstract_inverted_index.valuable | 168 |
| abstract_inverted_index.available | 87 |
| abstract_inverted_index.expanding | 178 |
| abstract_inverted_index.explored, | 46 |
| abstract_inverted_index.inference | 136 |
| abstract_inverted_index.potential | 144 |
| abstract_inverted_index.prompting | 48 |
| abstract_inverted_index.provided, | 119 |
| abstract_inverted_index.Resolution | 19 |
| abstract_inverted_index.approaches | 44 |
| abstract_inverted_index.complexity | 59 |
| abstract_inverted_index.effective, | 54 |
| abstract_inverted_index.highlights | 142 |
| abstract_inverted_index.in-context | 38 |
| abstract_inverted_index.languages. | 182 |
| abstract_inverted_index.supervised | 160 |
| abstract_inverted_index.trade-offs | 132 |
| abstract_inverted_index.Coreference | 18 |
| abstract_inverted_index.comparisons | 157 |
| abstract_inverted_index.coreference | 24, 151 |
| abstract_inverted_index.performance | 110 |
| abstract_inverted_index.referential | 70 |
| abstract_inverted_index.resolution, | 25 |
| abstract_inverted_index.submission. | 95 |
| abstract_inverted_index.traditional | 159 |
| abstract_inverted_index.Multilingual | 17 |
| abstract_inverted_index.availability | 65 |
| abstract_inverted_index.competition. | 77 |
| abstract_inverted_index.fine-tuning. | 154 |
| abstract_inverted_index.high-quality | 67 |
| abstract_inverted_index.long-context | 30, 89, 146 |
| abstract_inverted_index.low-resource | 181 |
| abstract_inverted_index.multilingual | 23, 150 |
| abstract_inverted_index.particularly | 55, 176 |
| abstract_inverted_index.closed-source | 88 |
| abstract_inverted_index.demonstrating | 108 |
| abstract_inverted_index.mini-testset, | 107 |
| abstract_inverted_index.relationships | 71 |
| abstract_inverted_index.significantly | 112 |
| abstract_inverted_index.task-specific | 153 |
| abstract_inverted_index.straightforward, | 164 |
| cited_by_percentile_year | |
| countries_distinct_count | 0 |
| institutions_distinct_count | 1 |
| citation_normalized_percentile |