Sample record. This work is fictional demonstration content. Results, people and institutions are placeholders; references in full-text articles are real publications.
Abstract
Users frequently mix languages within a single query, but retrieval benchmarks rarely reflect this. We construct code-mixed query sets for two language pairs and evaluate dense retrievers under different tokenisation and fine-tuning strategies. Adapting the query encoder with synthetic code-mixed data improves recall@20 by 9 points over a multilingual baseline, while transliteration normalisation helps for romanised queries but hurts for native-script queries.
Keywords dense retrieval · code-mixing · multilingual · tokenisation
Full text
CTM-PB-2026-0009-v1.pdf
Declarations
- Author contributions
- Author contributions are listed in the PDF.
- Funding
- Sample record – no funding to declare.
- Competing interests
- The authors declare no competing interests.
- Data and code availability
- Sample article. All results, names and affiliations in this demonstration article are synthetic placeholders.
Peer review
Open reports from independent reviewers. See how outcomes are determined.
- Status
- Under peer review
- Requested
- 7 Oct 2026
- Reviewers invited
- 3
- Report identity
- Signed reports
No reports yet
3 reviewers have been invited. Their reports will appear here as soon as they are submitted.
Versions
Version 1 · 5 October 2026Latest
Posted; peer review requested 7 October 2026.
10.xxxx/ctm.pb.2026.0009.v1
How to cite
Adebayo, O., Wagner, E., & Menon, R. (2026). Dense Retrieval for Code-Mixed Queries. CONTINUUM Publishing, CTM-PB-2026-0009. doi:10.xxxx/ctm.pb.2026.0009
© 2026 The authors. Published by AEX CONTINUUM under the CC BY 4.0 licence.