mlsanityData to 14 Oct 2026

Linguistic Nepotism: Trading-off Quality for Language Preference in Multilingual RAG

arXiv 2509.13930 ↗First posted 17 Sep 2025cs.CLcited by 7

On the map: Retrieval-augmented generation → Generation and context integration

Abstract (arXiv)

Multilingual Retrieval-Augmented Generation (mRAG) systems enable language models to answer knowledge-intensive queries with citation-supported responses across languages. Despite their growing use, an open questions is whether the mixture of different document languages impacts generation and citation behavior in unintended ways. To investigate this, we introduce a controlled methodology using model internals to measure language preference while holding other factors such as document relevance constant. Across eight languages and six open-weight models, we find that models preferentially cite English sources when queries are in English, with this bias amplified for lower-resource languages and for documents positioned mid-context. More crucially, we find that models sometimes trade-off document relevance for language preference, indicating that citation choices are not always driven by informativeness alone. Our findings shed light on how language models leverage multilingual context and influence citation behavior.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Placed in Generation and context integration, one of 15 directions in Retrieval-augmented generation. Builds on 61 of its 78 references that are in the map. First posted on arXiv 17 Sep 2025.

Builds on

61 of 78 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025The Cross-Lingual Cost: Retrieval Biases in RAG over Arabic-English Corpora10 Jul
  2. 2025XRAG: Cross-lingual Retrieval-Augmented Generation15 May
  3. 2025Multilingual Retrieval-Augmented Generation for Knowledge-Intensive Task4 Apr
  4. 2025On the Consistency of Multilingual Context Utilization in Retrieval-Augmented Generation1 Apr
  5. 2025Investigating Language Preference of Multilingual RAG Systems16 Feb
  6. 2024Not All Languages are Equal: Insights into Multilingual Retrieval-Augmented Generation29 Oct
  7. 2024Multilingual Retrieval Augmented Generation for Culturally-Sensitive Tasks: A Benchmark for Cross-lingual Robustness2 Oct
  8. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks22 May

Other references · 53

  1. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,824
  2. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  3. 2022Survey of Hallucination in Natural Language Generationcited by 2,075
  4. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,601
  5. 2020COMET: A Neural Framework for MT Evaluationcited by 1,033
  6. 2020Language-agnostic BERT Sentence Embeddingcited by 627
  7. 2023Enabling Large Language Models to Generate Text with Citationscited by 595
  8. 2019ELI5: Long Form Question Answeringcited by 575
  9. 2021A Dataset of Information-Seeking Questions and Answers Anchored in Research Paperscited by 486
  10. 2023Ragas: Automated Evaluation of Retrieval Augmented Generationcited by 418
  11. 2023Is ChatGPT Good at Search? Investigating Large Language Models as Re-Ranking Agentscited by 360
  12. 2023Benchmarking Large Language Models in Retrieval-Augmented Generationcited by 312
38 more
  1. 2023Evaluating Verifiability in Generative Search Enginescited by 303
  2. 2022Teaching language models to support answers with verified quotescited by 292
  3. 2025DeepResearcher: Scaling Deep Research via Reinforcement Learning in Real-world Environmentscited by 286
  4. 2024Do Llamas Work in English? On the Latent Language of Multilingual Transformerscited by 281
  5. 2023Fine-Tuning LLaMA for Multi-Stage Text Retrievalcited by 272
  6. 2023ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systemscited by 265
  7. 2021Measuring Attribution in Natural Language Generation Modelscited by 256
  8. 2022ASQA: Factoid Questions Meet Long-Form Answerscited by 234
  9. 2024Tower: An Open Multilingual Large Language Model for Translation-Related Taskscited by 206
  10. 2018Fast Lexically Constrained Decoding with Dynamic Beam Allocation for Neural Machine Translationcited by 182
  11. 2024Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Toolscited by 176
  12. 2021Hurdles to Progress in Long-form Question Answeringcited by 167
  13. 2024Long-form factuality in large language modelscited by 155
  14. 2025Deep Research Agents: A Systematic Examination And Roadmapcited by 128
  15. 2023Factuality Challenges in the Era of Large Language Modelscited by 126
  16. 2024Long-Context LLMs Meet RAG: Overcoming Challenges for Long Inputs in RAGcited by 123
  17. 2024ContextCite: Attributing Model Generation to Contextcited by 98
  18. 2024The Language Barrier: Dissecting Safety Challenges of LLMs in Multilingual Contextscited by 94
  19. 2023Document-Level Machine Translation with Large Language Modelscited by 90
  20. 2024LongCite: Enabling LLMs to Generate Fine-grained Citations in Long-context QAcited by 81
  21. 2025Do Multilingual LLMs Think In English?cited by 78
  22. 2021GooAQ: Open Question Answering with Diverse Answer Typescited by 53
  23. 2024LongRAG: A Dual-Perspective Retrieval-Augmented Generation Paradigm for Long-Context Question Answeringcited by 49
  24. 2023Understanding Retrieval Augmentation for Long-Form Question Answeringcited by 45
  25. 2025Lost in Multilinguality: Dissecting Cross-lingual Factual Inconsistency in Transformer Language Modelscited by 44
  26. 2024Beyond English-Centric LLMs: What Language Do Multilingual Language Models Think in?cited by 41
  27. 2024Long2RAG: Evaluating Long-Context & Long-Form Retrieval-Augmented Generation with Key Point Recallcited by 32
  28. 2024Retrieval-augmented generation in multilingual settingscited by 27
  29. 2024DelTA: An Online Document-Level Translation Agent Based on Multi-Level Memorycited by 27
  30. 2024Efficiently Exploring Large Language Models for Document-Level Machine Translation with In-context Learningcited by 23
  31. 2025SciArena: An Open Evaluation Platform for Non-Verifiable Scientific Literature-Grounded Taskscited by 20
  32. 2025Language Models Prefer What They Know: Relative Confidence Estimation via Confidence Preferencescited by 13
  33. 2022MIA 2022 Shared Task: Evaluating Cross-lingual Open-Retrieval Question Answering for 16 Diverse Languagescited by 11
  34. 2025The Translation Barrier Hypothesis: Multilingual Generation with Large Language Models Suffers from Implicit Translation Failurecited by 10
  35. 2025Controlled Retrieval-augmented Context Evaluation for Long-form RAGcited by 9
  36. 2024Faux Polyglot: A Study on Information Disparity in Multilingual Large Language Modelscited by 7
  37. 2024Multilingual Hallucination Gaps in Large Language Modelscited by 7
  38. 2024Language Fairness in Multilingual Information Retrievalcited by 5

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 6.

7 papers.

  1. 2025Exposing Citation Vulnerabilities in Generative Enginescited by 2
  2. 2026Enhancing Multilingual RAG Systems with Debiased Language Preference-Guided Query Fusioncited by 2
  3. 2026GSM-Plus-BN: A Perturbation-Based Benchmark for Bangla Mathematical Reasoning in Large Language Modelscited by 2
  4. 2026LAMAR: An Open Language-Aware Multilingual Alignment Rerankercited by 0
  5. 2026CliniCIRCA: A Modular LLM Framework for Constructing Longitudinal Mental Health Patient Journeys from Raw EHR Narrativescited by 0
  6. 2026ChemCLIR-Bench: Benchmarking Cross-Lingual Information Retrieval in Multilingual Chemical Patentscited by 0
  7. 2026Where's Waldo? Query-language Preference under Cross-lingual Knowledge Disparitiescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026All Languages Matter: Understanding and Mitigating Language Bias in Multilingual RAG
  2. 2024Large Language Models Reflect Human Citation Patterns with a Heightened Citation Bias
  3. 2024On the Capacity of Citation Generation by Large Language Models
  4. 2026How Do LLMs Cite? A Mechanistic Interpretation of Attribution in Retrieval-Augmented Generation
  5. 2026Aligning Large Language Model Behavior with Human Citation Preferences
  6. 2024Enhancing Multilingual Information Retrieval in Mixed Human Resources Environments: A RAG Model Implementation for Multicultural Enterprise
  7. 2025Language Drift in Multilingual Retrieval-Augmented Generation: Characterization and Decoding-Time Mitigation
  8. 2026On the Role of Citations in Preference Data