Elamur Research⌕
☰
home / papers / 2026 / Can you map it to English? The Role of Cross-Lingual Alignment in the Multilingual Performance of LLMs

Can you map it to English? The Role of Cross-Lingual Alignment in the Multilingual Performance of LLMs

Paper · 2026 · Kartik Ravisankar, HyoJung Han, Sarah Wiegreffe et al.

Summary

Decoder-only LLMs are trained mostly on English text, yet they can handle many other languages. It remains unclear why a given non-English prediction succeeds or fails. Earlier work such as MEXA linked cross-lingual alignment to accuracy only for whole languages, and it measured alignment on task-independent parallel text. This paper moves the analysis down to individual test items.

The authors propose three per-layer binary scores based on cosine similarity between non-English inputs and their English parallels:

  • the Discriminative Alignment Index (DALI),
  • a stricter variant, DALI st,
  • a task-specific MEXA T.

They apply these scores to 24 non-English languages and three multiple-choice NLU benchmarks. Items are grouped into Transfer Success (correct in both English and the target language) and Transfer Failure (correct only in English). Success items are consistently more aligned with English, especially in middle layers.

Activation patching then tests causality. Inserting the matching English activations into middle layers repairs failures more often than control patches do. This supports a causal role for alignment and motivates inference-time steering for weaker languages.

Key points

  • How DALI works: DALI scores 1 when cross-lingual matched premise+option pairs are more similar than mismatched cross-lingual pairs. DALI st adds a second condition: matched pairs must also beat mismatched pairs within the same language. This reduces false positives, since a 2-option item yields only two cross-lingual mismatches.
  • Why binary scores: Scores are binary rather than raw cosine values because of representation anisotropy. This follows MEXA (MEXA F), which uses FLORES-200 and also the Bible corpus.
  • Patching method and results: English activations at layer λ are patched into the forward pass of Transfer Failure items. Semantically equivalent patches outperform non-equivalent controls, and the effect is concentrated in specific intermediate layers.
  • Confidence effect: In a Belebele French analysis, flips caused by aligned patches show lower output entropy, meaning more confident predictions.
  • Limitations and code: The study covers only discriminative tasks, uses human-translated benchmarks that may contain translation artifacts, and examines alignment with English only. Code is available at github.com/Kartik21/XLingAlignment.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓University of Maryland, College ParkCollege Park · United States132 works · 25,380 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 2Research subjects and languages the work is about.Language models & pretraining · 3,320 worksMultilingual & cross-lingual NLP · 828 works

Cite

Cite as Kartik Ravisankar, HyoJung Han, Sarah Wiegreffe and Marine Carpuat. 2026. Can you map it to English? The Role of Cross-Lingual Alignment in the Multilingual Performance of LLMs. Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.eacl-long.225

@inproceedings{ravisankar2026can,
  title = {{Can you map it to English? The Role of Cross-Lingual Alignment in the Multilingual Performance of LLMs}},
  author = {Kartik Ravisankar, HyoJung Han, Sarah Wiegreffe and Marine Carpuat},
  booktitle = {Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.eacl-long.225},
  url = {https://doi.org/10.18653/v1/2026.eacl-long.225}
}

References

39 references, 23 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Orevaoghene Ahia, Sachin Kumar, Hila Gonen, Jungo Kasai, David Mortensen, Noah Smith, and Yulia Tsvetkov. 2023. Do all languages cost the same? tokenization in the era of commercial language models. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 9904…
  2. Mikel Artetxe, Gorka Labaka, and Eneko Agirre. 2020. Translation artifacts in cross-lingual transfer learning. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 7674–7684, Online. Association for Computational Linguistics.
  3. Viraat Aryabumi, John Dang, Dwarak Talupuru, Saurabh Dash, David Cairuz, Hangyu Lin, Bharat Venkitesh, Madeline Smith, Jon Ander Campos, Yi Chern Tan, et al. 2024. Aya 23: Open weight releases to further multilingual progress.
  4. Niyati Bafna, Tianjian Li, Kenton Murray, David R. Mortensen, David Yarowsky, Hale Sirin, and Daniel Khashabi. 2025. The translation barrier hypothesis: Multilingual generation with large language models suffers from implicit translation failure.
  5. Lucas Bandarkar, Davis Liang, Benjamin Muller, Mikel Artetxe, Satya Narayan Shukla, Donald Husa, Naman Goyal, Abhinandan Krishnan, Luke Zettlemoyer, and Madian Khabsa. 2024. The belebele benchmark: a parallel reading comprehension dataset in 122 language variants. In Proceedings of the 62nd Annual M…
  6. Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer, and Veselin Stoyanov. 2020. Unsupervised cross-lingual representation learning at scale. In Proceedings of the 58th Annual Meeting of the Association fo… find
  7. Alexis Conneau, Ruty Rinott, Guillaume Lample, Adina Williams, Samuel Bowman, Holger Schwenk, and Veselin Stoyanov. 2018. XNLI: Evaluating cross-lingual sentence representations. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 2475–2485, Brussels, Bel… find
  8. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,…
  9. Clément Dumas, Chris Wendler, Veniamin Veselovsky, Giovanni Monea, and Robert West. 2025. Separating tongue from thought: Activation patching reveals language-agnostic concept representations in transformers. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics…
  10. Kawin Ethayarajh. 2019. How contextual are contextualized word representations? Comparing the geometry of BERT, ELMo, and GPT-2 embeddings. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Process… find
  11. Julen Etxaniz, Gorka Azkune, Aitor Soroa, Oier Lopez de Lacalle, and Mikel Artetxe. 2023. Do multilingual language models think better in english?
  12. Jaden Fiotto-Kaufman, Alexander R. Loftus, Eric Todd, Jannik Brinkmann, Koyena Pal, Dmitrii Troitskii, Michael Ripa, Adam Belfki, Can Rager, Caden Juang, et al. 2025. Nnsight and ndif: Democratizing access to open-weight foundation model internals. find
  13. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, et al. 2024. The llama 3 herd of models.
  14. Katharina Hämmerl, Jindřich Libovický, and Alexander Fraser. 2024. Understanding cross-lingual Alignment—A survey. In Findings of the Association for Computational Linguistics: ACL 2024, pages 10922–10943, Bangkok, Thailand. Association for Computational Linguistics. find
  15. Amir Hossein Kargaran, Ali Modarressi, Nafiseh Nikeghbal, Jana Diesner, François Yvon, and Hinrich Schuetze. 2025. MEXA: Multilingual evaluation of English-centric LLMs via cross-lingual alignment. In Findings of the Association for Computational Linguistics: ACL 2025, pages 27001–27023, Vienna, Aus… find
  16. Chong Li, Shaonan Wang, Jiajun Zhang, and Chengqing Zong. 2024. Improving in-context learning of multilingual generative language models with cross-lingual alignment. find
  17. Zihao Li, Yucheng Shi, Zirui Liu, Fan Yang, Ali Payani, Ninghao Liu, and Mengnan Du. 2025. Language ranker: a metric for quantifying llm performance across high and low-resource languages. In Proceedings of the Thirty-Ninth AAAI Conference on Artificial Intelligence and Thirty-Seventh Conference on…
  18. Zheng Wei Lim, Alham Fikri Aji, and Trevor Cohn. 2025. Language-specific latent process hinders cross-lingual performance. find
  19. Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, et al. 2022. Few-shot learning with multilingual generative language models. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Proc…
  20. Danni Liu and Jan Niehues. 2025. Middle-layer representation alignment for cross-lingual transfer in fine-tuned LLMs. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 15979– 15996, Vienna, Austria. Association for Computational…
  21. Meng Lu, Ruochen Zhang, Carsten Eickhoff, and Ellie Pavlick. 2025. Paths not taken: Understanding and mending the multilingual factual recall pipeline. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 15077–15107, Suzhou, China. Association for Computa…
  22. Thomas Mayer and Michael Cysouw. 2014. Creating a massively parallel Bible corpus. In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), pages 3158– 3163, Reykjavik, Iceland. European Language Resources Association (ELRA).
  23. Kevin Meng, David Bau, Alex J Andonian, and Yonatan Belinkov. 2022. Locating afnd editing factual associations in GPT. In Advances in Neural Information Processing Systems. find
  24. Aaron Mueller, Jannik Brinkmann, Millicent L. Li, Samuel Marks, Koyena Pal, Nikhil Prakash, Can Rager, Aruna Sankaranarayanan, Arnab Sen Sharma, Jiuding Sun, Eric Todd, David Bau, and Yonatan Belinkov. 2024. The quest for the right mediator: A history, survey, and theoretical grounding of causal int… — [arXiv:2408.01416](https://arxiv.org/abs/2408.01416) find
  25. Niklas Muennighoff, Thomas Wang, Lintang Sutawika, Adam Roberts, Stella Biderman, Teven Le Scao, M Saiful Bari, Sheng Shen, Zheng Xin Yong, Hailey Schoelkopf, et al. 2023. Crosslingual generalization through multitask finetuning. In Proceedings of the 61st Annual Meeting of the Association for Compu…
  26. Arvind Neelakantan, Tao Xu, Raul Puri, Alec Radford, Jesse Michael Han, Jerry Tworek, Qiming Yuan, Nikolas Tezak, Jong Wook Kim, Chris Hallacy, et al. 2022. Text and code embeddings by contrastive pre-training. nosalgebraist. 2020. Interpreting gpt: The logit lens; accessed: 2025-09-25.
  27. Edoardo Maria Ponti, Goran Glavaš, Olga Majewska, Qianchu Liu, Ivan Vulić, and Anna Korhonen. 2020. XCOPA: A multilingual dataset for causal commonsense reasoning. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 2362–2376, Online. Association…
  28. Anirudh Sundar, Sinead Williamson, Katherine Metcalf, Barry-John Theobald, Skyler Seto, and Masha Fedzechkina. 2025. Steering into new embedding spaces: Analyzing cross-lingual alignment induced by model interventions in multilingual language models. In Proceedings of the 63rd Annual Meeting of the…
  29. NLLB Team, Marta R. Costa-jussà, James Cross, Onur Çelebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, et al. 2022. No language left behind: Scaling human-centered machine translation. find
  30. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. 2023. Llama 2: Open foundation and fine-tuned chat models.
  31. Jesse Vig, Sebastian Gehrmann, Yonatan Belinkov, Sharon Qian, Daniel Nevo, Yaron Singer, and Stuart Shieber. 2020. Investigating gender bias in language models using causal mediation analysis. In Advances in Neural Information Processing Systems, volume 33, pages 12388–12401. find
  32. Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, and Furu Wei. 2024. Improving text embeddings with large language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 11897–11916, Bangkok, Thailand. Asso…
  33. Mingyang Wang, Heike Adel, Lukas Lange, Yihong Liu, Ercong Nie, Jannik Strötgen, and Hinrich Schuetze. 2025a. Lost in multilinguality: Dissecting cross-lingual factual inconsistency in transformer language models. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguis…
  34. Weixuan Wang, Minghao Wu, Barry Haddow, and Alexandra Birch. 2025b. Bridging the language gaps in large language models with inference-time cross-lingual intervention. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5418–5433,… find
  35. Chris Wendler, Veniamin Veselovsky, Giovanni Monea, and Robert West. 2024. Do llamas work in English? on the latent language of multilingual transformers. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 15366–15394, Bangkok, T…
  36. Sarah Wiegreffe, Oyvind Tafjord, Yonatan Belinkov, Hannaneh Hajishirzi, and Ashish Sabharwal. 2025. Answer, assemble, ace: Understanding how LMs answer multiple choice questions. In The Thirteenth International Conference on Learning Representations. find
  37. Bryan Wilie, Samuel Cahyawijaya, Junxian He, and Pascale Fung. 2025. High-dimensional interlingual representations of large language models. In Proceedings of the 7th Workshop on Research in Computational Linguistic Typology and Multilingual NLP, pages 122–155, Vienna, Austria. Association for Compu… find
  38. Yinfei Yang, Yuan Zhang, Chris Tar, and Jason Baldridge. 2019. PAWS-X: A cross-lingual adversarial dataset for paraphrase identification. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processin…
  39. Shaolei Zhang, Qingkai Fang, Zhuocheng Zhang, Zhengrui Ma, Yan Zhou, Langlin Huang, Mengyu Bu, Shangtong Gui, Yunji Chen, Xilin Chen, and Yang Feng. 2023. Bayling: Bridging cross-lingual alignment and instruction following through interactive translation for large language models. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ravisankar-2026-can-you-map-it-english/