mlsanityData to 14 Oct 2026

MERLIN: Multi-Stage Curriculum Alignment for Multilingual Encoder-LLM Integration in Cross-Lingual Reasoning

arXiv 2509.08105 ↗First posted 9 Sep 2025cs.CLcited by 2

On the map: LLM reasoning → Reasoning benchmarks and evaluation

Abstract (arXiv)

Large language models excel in English but still struggle with complex reasoning in many low-resource languages (LRLs). Existing encoder-plus-decoder methods such as LangBridge and MindMerger raise accuracy on mid and high-resource languages, yet they leave a large gap on LRLs. We present MERLIN, a two-stage model-stacking framework that applies a curriculum learning strategy -- from general bilingual bitext to task-specific data -- and adapts only a small set of DoRA weights. On the AfriMGSM benchmark MERLIN improves exact-match accuracy by +12.9 pp over MindMerger and outperforms GPT-4o-mini. It also yields consistent gains on MGSM and MSVAMP (+0.9 and +2.8 pp), demonstrating effectiveness across both low and high-resource settings.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 40 of its 61 references that are in the map. First posted on arXiv 9 Sep 2025.

Builds on

40 of 61 references.

Closest ancestors · 5

Same direction, newest first

  1. 2025LayAlign: Enhancing Multilingual Reasoning in Large Language Models via Layer-Wise Adaptive Fusion and Alignment Strategy17 Feb
  2. 2025SLAM: Towards Efficient Multilingual Reasoning via Selective Language Alignment7 Jan
  3. 2024MindMerger: Efficient Boosting LLM Reasoning in non-English Languages27 May
  4. 2024Multilingual Large Language Model: A Survey of Resources, Taxonomy and Frontiers7 Apr
  5. 2024Question Translation Training for Better Multilingual Reasoning15 Jan

Other references · 35

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2024The Llama 3 Herd of Modelscited by 14,715
  3. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  4. 2023QLoRA: Efficient Finetuning of Quantized LLMscited by 3,418
  5. 2024Gemini 1.5: Unlocking multimodal understanding across millions of tokens of contextcited by 3,224
  6. 2017A Broad-Coverage Challenge Corpus for Sentence Understanding through Inferencecited by 2,682
  7. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,998
  8. 2024Mixtral of Expertscited by 1,823
  9. 2022No Language Left Behind: Scaling Human-Centered Machine Translationcited by 1,312
  10. 2024DoRA: Weight-Decomposed Low-Rank Adaptationcited by 758
  11. 2023MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Modelscited by 735
  12. 2024Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Surveycited by 656
23 more
  1. 2022Language Models are Multilingual Chain-of-Thought Reasonerscited by 530
  2. 2024Aya Model: An Instruction Finetuned Open-Access Multilingual Language Modelcited by 295
  3. 2024Aya 23: Open Weight Releases to Further Multilingual Progresscited by 123
  4. 2020Low-resource Languages: A Review of Past Work and Future Challengescited by 91
  5. 2022NaijaSenti: A Nigerian Twitter Sentiment Corpus for Multilingual Sentiment Analysiscited by 64
  6. 2023Breaking Language Barriers in Multilingual Mathematical Reasoning: Insights and Observationscited by 59
  7. 2024Adapting Large Language Models for Document-Level Machine Translationcited by 59
  8. 2023AfroBench: How Good are Large Language Models on African Languages?cited by 43
  9. 2024IrokoBench: A New Benchmark for African Languages in the Age of Large Language Modelscited by 43
  10. 2025Middle-Layer Representation Alignment for Cross-Lingual Transfer in Fine-Tuned LLMscited by 29
  11. 2023Efficiently Adapting Pretrained Language Models To New Languagescited by 28
  12. 2024Chinese Tiny LLM: Pretraining a Chinese-Centric Large Language Modelcited by 25
  13. 2025SEA-LION: Southeast Asian Languages in One Networkcited by 22
  14. 2024LexC-Gen: Generating Data for Extremely Low-Resource Languages with Large Language Models and Bilingual Lexiconscited by 11
  15. 2025Lugha-Llama: Adapting Large Language Models for African Languagescited by 8
  16. 2024Low-resource neural machine translation with morphological modelingcited by 7
  17. 2023A Challenging Benchmark for Low-Resource Learningcited by 5
  18. 2024A Systematic Review of NLP for Ghanaian Languages: Datasets, Models, and a Research Roadmapcited by 5
  19. 2024Chain-of-Translation Prompting (CoTR): A Novel Prompting Technique for Low Resource Languagescited by 5
  20. 2022Text normalization for low-resource languages: the case of Liguriancited by 4
  21. 2024A Morphology-Based Investigation of Positional Encodingscited by 4
  22. 2024Constructing and Expanding Low-Resource and Underrepresented Parallel Datasets for Indonesian Local Languagescited by 3
  23. 2025Improving Multilingual Math Reasoning for African Languagescited by 1

Built on by

Cited by year (2026 to 14 Oct)

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

2 papers.

  1. 2025Opportunities and Challenges of Natural Language Processing for Low-Resource Senegalese Languages in Social Science Researchcited by 0
  2. 2026Building Multilingual Bridges: Data Mixing as the Pillar of Generalization for In-Language Reasoningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024LinguaLIFT: An Effective Two-stage Instruction Tuning Framework for Low-Resource Language Reasoning
  2. 2024xCoT: Cross-lingual Instruction Tuning for Cross-lingual Chain-of-Thought Reasoning
  3. 2025Gradual Code-Switching as Inference-Time Cross-Lingual Representational Alignment for LLMs
  4. 2025Cross-Lingual Consistency: A Novel Inference Framework for Advancing Reasoning in Large Language Models
  5. 2025Reasoning Transfer for an Extremely Low-Resource and Endangered Language: Bridging Languages Through Sample-Efficient Language Understanding
  6. 2025Unlocking Multilingual Reasoning Capability of LLMs and LVLMs through Representation Engineering
  7. 2026Rethinking the Multilingual Reasoning Gap with Layer Swap
  8. 2025IRLBench: A Multi-modal, Culturally Grounded, Parallel Irish-English Benchmark for Open-Ended LLM Reasoning Evaluation