Elamur Research⌕
☰
home / papers / 2026 / Steering Large Language Models for Machine Translation Personalization

Steering Large Language Models for Machine Translation Personalization

Paper · 2026 · Daniel Scalena, Gabriele Sarti, Arianna Bisazza et al.

Summary

The paper asks whether an LLM can produce translations in the style of a specific human translator when only a handful of that translator's examples are available. It studies this in literary translation, where individual stylistic choices are most visible. The authors use the PAR 3 dataset, which pairs novels in seven languages with two professional English translations each. They first confirm that translator styles can be told apart automatically and that this style information is encoded in intermediate model layers. They then compare prompting strategies with inference-time steering methods on Llama 3.1 8B Instruct and Gemma 2 (2B and 9B). Their main method is contrastive steering, which amplifies style-related sparse autoencoder (SAE) latents during generation. This approach matches or beats multi-shot prompting on style accuracy without lowering translation quality. It also avoids the extra inference cost of long in-context prompts. Probing suggests that SAE steering and prompting change model activations in comparable ways. The work matters because it offers a cheap, scalable route to personalized MT and links prompting behavior to interpretable model features.

Key points

  • Data: PAR 3 covers one novel per source language (IT, FR, NL, DE, RU, JA, ZH) translated into English, for example Pinocchio, Crime and Punishment and Dream of the Red Chamber. Each example is a triple of MT output and two human translations (H1, H2), with 20 extra examples per language set aside for prompting and steering.
  • Style detection: XLM-based classifiers distinguished H1, H2 and MT output with 77% accuracy (Japanese) up to 99% (Chinese), averaging 86%.
  • Prompting baseline: Multi-shot prompting with 20 examples raised classifier-measured personalization from roughly 0.08–0.10 to 0.24–0.32. COMET scores stayed around 0.69–0.73.
  • Methods compared:
    • Prompting: zero-shot, multi-shot and explanation-based prompts.
    • Steering: ActAdd, ReFT and contrastive SAE steering ("SAE Cont.").
  • Mechanistic link: Probes trained only on multi-shot prompt activations could predict whether SAE steering succeeded for the Gemma models. This suggests the selected SAE features act like compressed summaries of the prompt demonstrations.

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 GroningenGroningen · Netherlands70 works · 5,142 citations🎓Northeastern UniversityBoston · United States57 works · 6,533 citations🎓University of Milano-BicoccaMilan · Italy15 works · 4,946 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 1Research subjects and languages the work is about.Machine translation · 1,087 works

Cite

Cite as Daniel Scalena, Gabriele Sarti, Arianna Bisazza, Elisabetta Fersini and Malvina Nissim. 2026. Steering Large Language Models for Machine Translation Personalization. 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.217

@inproceedings{scalena2026steering,
  title = {{Steering Large Language Models for Machine Translation Personalization}},
  author = {Daniel Scalena, Gabriele Sarti, Arianna Bisazza, Elisabetta Fersini and Malvina Nissim},
  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.217},
  url = {https://doi.org/10.18653/v1/2026.eacl-long.217}
}

References

60 references, 19 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Andy Arditi, Oscar Obeso, Aaquib Syed, Daniel Paleka, Nina Panickssery, Wes Gurnee, and Neel Nanda. 2024. Refusal in language models is mediated by a single direction. In Advances in Neural Information Processing Systems, volume 37, pages 136037– 136083. Curran Associates, Inc.
  2. Yonatan Belinkov. 2022. Probing classifiers: Promises, shortcomings, and advances. Computational Linguistics, 48(1):207–219.
  3. Nathaniel Berger, Stefan Riezler, Miriam Exel, and Matthias Huck. 2024. Post-edits are preferences too. In Proceedings of the Ninth Conference on Machine Translation, pages 1289–1300, Miami, Florida, USA. Association for Computational Linguistics. find
  4. Eleftheria Briakou, Jiaming Luo, Colin Cherry, and Markus Freitag. 2024. Translating step-by-step: Decomposing the translation process for improved translation quality of long-form texts. In Proceedings of the Ninth Conference on Machine Translation, pages 1301–1317, Miami, Florida, USA. Association…
  5. Sviatoslav Chalnev, Matthew Siu, and Arthur Conmy. 2024. Improving steering vectors by targeting sparse autoencoder features. find
  6. David Chanin, Anthony Hunter, and Oana-Maria Camburu. 2024. Identifying linear relational concepts in large language models. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), page… find
  7. 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
  8. Hoagy Cunningham, Aidan Ewart, Logan Riggs, Robert Huben, and Lee Sharkey. 2023. Sparse autoencoders find highly interpretable features in language models. find
  9. Javier Ferrando, Oscar Obeso, Senthooran Rajamanoharan, and Neel Nanda. 2025. Do i know this entity? knowledge awareness and hallucinations in language models. find
  10. Jaden Fiotto-Kaufman, Alexander R Loftus, Eric Todd, Jannik Brinkmann, Caden Juang, Koyena Pal, Can Rager, Aaron Mueller, Samuel Marks, Arnab Sen Sharma, Francesca Lucchetti, Michael Ripa, Adam Belfki, Nikhil Prakash, Sumeet Multani, Carla Brodley, Arjun Guha, Jonathan Bell, Byron Wallace, and David… find
  11. Lucie Flekova, Jordan Carpenter, Salvatore Giorgi, Lyle Ungar, and Daniel Preoţiuc-Pietro. 2016. Analyzing biases in human perception of user age and gender from text. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 843–854,… find
  12. Xavier Garcia and Orhan Firat. 2022. Using natural language prompts for machine translation. find
  13. Ana Guerberof-Arenas and Antonio Toral. 2022. Creativity in translation. Translation Spaces, 11(2):184– 212. find
  14. Roee Hendel, Mor Geva, and Amir Globerson. 2023. In-context learning creates task vectors. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 9318–9333, Singapore. Association for Computational Linguistics.
  15. Marzena Karpinska and Mohit Iyyer. 2023. Large language models effectively leverage document-level context for literary translation, but critical errors persist. In Proceedings of the Eighth Conference on Machine Translation, pages 419–451, Singapore. Association for Computational Linguistics.
  16. Jihyeon Lee, Taehee Kim, Yunwon Tae, Cheonbok Park, and Jaegul Choo. 2023. PePe: Personalized post-editing model utilizing user-generated post-edits. In Findings of the Association for Computational Linguistics: EACL 2023, pages 239–253, Dubrovnik, Croatia. Association for Computational Linguistics. find
  17. Kenneth Li, Oam Patel, Fernanda Viégas, Hanspeter Pfister, and Martin Wattenberg. 2023. Inference-time intervention: Eliciting truthful answers from a language model. In Advances in Neural Information Processing Systems, volume 36, pages 41451–41530. Curran Associates, Inc.
  18. Tom Lieberum, Senthooran Rajamanoharan, Arthur Conmy, Lewis Smith, Nicolas Sonnerat, Vikrant Varma, Janos Kramar, Anca Dragan, Rohin Shah, and Neel Nanda. 2024. Gemma scope: Open sparse autoencoders everywhere all at once on gemma 2. In Proceedings of the 7th BlackboxNLP Workshop: Analyzing and Inte… find
  19. Huan Lin, Liang Yao, Baosong Yang, Dayiheng Liu, Haibo Zhang, Weihua Luo, Degen Huang, and Jinsong Su. 2021. Towards user-driven neural machine translation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natur… find
  20. Xiaoming Liu, Zhaohan Zhang, Yichen Wang, Hang Pu, Yu Lan, and Chao Shen. 2023. CoCo: Coherence-enhanced machine-generated text detection under low resource with contrastive learning. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 16167–16188, Singap… find
  21. Samuel Marks, Can Rager, Eric J Michaud, Yonatan Belinkov, David Bau, and Aaron Mueller. 2025. Sparse feature circuits: Discovering and editing interpretable causal graphs in language models. In The Thirteenth International Conference on Learning Representations. find
  22. Evgeny Matusov. 2019. The challenges of using neural machine translation for literature. In Proceedings of the Qualities of Literary Machine Translation, pages 10–19, Dublin, Ireland. European Association for Machine Translation. find
  23. Thomas McGrath, Daniel Balsam, Myra Deng, and Eric Ho. 2024. Understanding and steering llama 3 with sparse autoencoders. find
  24. Paul Michel and Graham Neubig. 2018. Extreme adaptation for personalized neural machine translation. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 312–318, Melbourne, Australia. Association for Computational Linguistics.
  25. Tomas Mikolov, Wen-tau Yih, and Geoffrey Zweig. 2013. Linguistic regularities in continuous space word representations. In Proceedings of the 2013 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 746–751, Atlanta, Georgia.…
  26. Joss Moorkens, Antonio Toral, Sheila Castilho, and Andy Way. 2018. Translators’perceptions of literary post-editing using statistical and neural machine translation. Translation Spaces, 7(2):240–262. find
  27. Maria Nadejde, Anna Currey, Benjamin Hsu, Xing Niu, Marcello Federico, and Georgiana Dinu. 2022. CoCoA-MT: A dataset and benchmark for contrastive controlled MT with application to formality. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 616–632, Seattle, United Sta… find
  28. Xing Niu, Marianna Martindale, and Marine Carpuat. 2017. A study of style in machine translation: Controlling the formality of machine translation output. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 2814–2819, Copenhagen, Denmark. Association for… find
  29. OpenAI. 2024. Gpt-4 technical report.
  30. Ella Rabinovich, Raj Nath Patel, Shachar Mirkin, Lucia Specia, and Shuly Wintner. 2017. Personalized machine translation: Preserving original author traits. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers, pages 10…
  31. Ricardo Rei, Craig Stewart, Ana C Farinha, and Alon Lavie. 2020. COMET: A neural framework for MT evaluation. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 2685–2702, Online. Association for Computational Linguistics.
  32. Nina Rimsky, Nick Gabrieli, Julian Schulz, Meg Tong, Evan Hubinger, and Alexander Turner. 2024. Steering llama 2 via contrastive activation addition. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 15504–15522, Bangkok, Thaila… find
  33. Gabriele Sarti, Phu Mon Htut, Xing Niu, Benjamin Hsu, Anna Currey, Georgiana Dinu, and Maria Nadejde. 2023. RAMP: Retrieval and attribute-marking enhanced prompting for attribute-controlled translation. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volum… find
  34. Danielle Saunders and Bill Byrne. 2020. Reducing gender bias in neural machine translation as a domain adaptation problem. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 7724–7736, Online. Association for Computational Linguistics.
  35. Daniel Scalena, Gabriele Sarti, and Malvina Nissim. 2024. Multi-property steering of large language models with dynamic activation composition. In Proceedings of the 7th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP, pages 577–603, Miami, Florida, US. Association for Compu… find
  36. Rico Sennrich, Barry Haddow, and Alexandra Birch. 2016. Controlling politeness in neural machine translation via side constraints. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 35–40, San Diego…
  37. Gemma Team. 2024a. Gemma 2: Improving open language models at a practical size. find
  38. Llama Team. 2024b. The llama 3 herd of models.
  39. Adly Templeton, Tom Conerly, Jonathan Marcus, Jack Lindsey, Trenton Bricken, Brian Chen, Adam Pearce, Craig Citro, Emmanuel Ameisen, Andy Jones, Hoagy Cunningham, Nicholas L Turner, Callum McDougall, Monte MacDiarmid, C. Daniel Freeman, Theodore R. Sumers, Edward Rees, Joshua Batson, Adam Jermyn, Sh… find
  40. Katherine Thai, Marzena Karpinska, Kalpesh Krishna, Bill Ray, Moira Inghilleri, John Wieting, and Mohit Iyyer. 2022. Exploring document-level literary machine translation with parallel paragraphs from world literature. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Pr… find
  41. Eric Todd, Millicent L. Li, Arnab Sen Sharma, Aaron Mueller, Byron C. Wallace, and David Bau. 2024. Function vectors in large language models. In Proceedings of the 2024 International Conference on Learning Representations. ArXiv:2310.15213. find
  42. Antonio Toral and Andy Way. 2015. Translating literary text between related languages using SMT. In Proceedings of the Fourth Workshop on Computational Linguistics for Literature, pages 123–132, Denver, Colorado, USA. Association for Computational Linguistics. find
  43. Antonio Toral and Andy Way. 2018. What Level of Quality Can Neural Machine Translation Attain on Literary Text?, pages 263–287. Springer International Publishing, Cham. find
  44. Alexander Matt Turner, Lisa Thiergart, Gavin Leech, David Udell, Juan J. Vazquez, Ulisse Mini, and Monte MacDiarmid. 2024. Steering language models with activation engineering.
  45. Eva Vanmassenhove, Christian Hardmeier, and Andy Way. 2018. Getting gender right in neural machine translation. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 3003–3008, Brussels, Belgium. Association for Computational Linguistics.
  46. Rob Voigt and Dan Jurafsky. 2012. Towards a literary machine translation: The role of referential cohesion. In Proceedings of the NAACL-HLT 2012 Workshop on Computational Linguistics for Literature, pages 18–25, Montréal, Canada. Association for Computational Linguistics. find
  47. Longyue Wang, Siyou Liu, Chenyang Lyu, Wenxiang Jiao, Xing Wang, Jiahao Xu, Zhaopeng Tu, Yan Gu, Weiyu Chen, Minghao Wu, Liting Zhou, Philipp Koehn, Andy Way, and Yulin Yuan. 2024a. Findings of the WMT 2024 shared task on discourse-level literary translation. In Proceedings of the Ninth Conference o… find
  48. Longyue Wang, Chenyang Lyu, Tianbo Ji, Zhirui Zhang, Dian Yu, Shuming Shi, and Zhaopeng Tu. 2023a. Document-level machine translation with large language models. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 16646–16661, Singapore. Association for C…
  49. Longyue Wang, Zhaopeng Tu, Yan Gu, Siyou Liu, Dian Yu, Qingsong Ma, Chenyang Lyu, Liting Zhou, Chao-Hong Liu, Yufeng Ma, Weiyu Chen, Yvette Graham, Bonnie Webber, Philipp Koehn, Andy Way, Yulin Yuan, and Shuming Shi. 2023b. Findings of the WMT 2023 shared task on discourse-level literary translation… find
  50. Yue Wang, Cuong Hoang, and Marcello Federico. 2021. Towards modeling the style of translators in neural machine translation. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1193–1199, Online. Ass… find
  51. Yuxia Wang, Jonibek Mansurov, Petar Ivanov, Jinyan Su, Artem Shelmanov, Akim Tsvigun, Osama Mohammed Afzal, Tarek Mahmoud, Giovanni Puccetti, and Thomas Arnold. 2024b. SemEval-2024 task 8: Multidomain, multimodel and multilingual machine-generated text detection. In Proceedings of the 18th Internati… find
  52. Yuxia Wang, Jonibek Mansurov, Petar Ivanov, Jinyan Su, Artem Shelmanov, Akim Tsvigun, Osama Mohammed Afzal, Tarek Mahmoud, Giovanni Puccetti, Thomas Arnold, Alham Aji, Nizar Habash, Iryna Gurevych, and Preslav Nakov. 2024c. M4GTbench: Evaluation benchmark for black-box machine-generated text detecti… find
  53. Minghao Wu, Jiahao Xu, Yulin Yuan, Gholamreza Haffari, Longyue Wang, Weihua Luo, and Kaifu Zhang. 2025a. (Perhaps) beyond human translation: Harnessing multi-agent collaboration for translating ultra-long literary texts.
  54. Zhengxuan Wu, Aryaman Arora, Atticus Geiger, Zheng Wang, Jing Huang, Dan Jurafsky, Christopher D. Manning, and Christopher Potts. 2025b. Axbench: Steering llms? even simple baselines outperform sparse autoencoders. find
  55. Zhengxuan Wu, Aryaman Arora, Zheng Wang, Atticus Geiger, Dan Jurafsky, Christopher D Manning, and Christopher Potts. 2024. ReFT: Representation fine-tuning for language models. In The Thirty-eighth Annual Conference on Neural Information Processing Systems. find
  56. Wu Youyou, Michal Kosinski, and David Stillwell. 2015. Computer-based personality judgments are more accurate than those made by humans. Proceedings of the National Academy of Sciences, 112(4):1036– 1040. find
  57. Zeyu Yun, Yubei Chen, Bruno Olshausen, and Yann Le-Cun. 2021. Transformer visualization via dictionary learning: contextualized embedding as a linear superposition of transformer factors. In Proceedings of Deep Learning Inside Out (DeeLIO): The 2nd Workshop on Knowledge Extraction and Integration fo… find
  58. Peng Zhang, Zhengqing Guan, Baoxi Liu, Xianghua (Sharon) Ding, Tun Lu, Hansu Gu, and Ning Gu. 2022. Building user-oriented personalized machine translator based on user-generated textual content. Proc. ACM Hum.-Comput. Interact., 6(CSCW2). find
  59. Yu Zhao, Alessio Devoto, Giwon Hong, Xiaotang Du, Aryo Pradipta Gema, Hongru Wang, Xuanli He, Kam-Fai Wong, and Pasquale Minervini. 2025. Steering knowledge selection behaviours in LLMs via SAE-based representation engineering. In Proceedings of the 2025 Conference of the Nations of the Americas Cha… find
  60. Andy Zou, Long Phan, Sarah Chen, James Campbell, Phillip Guo, Richard Ren, Alexander Pan, Xuwang Yin, Mantas Mazeika, Ann-Kathrin Dombrowski, Shashwat Goel, Nathaniel Li, Michael J. Byun, Zifan Wang, Alex Mallen, Steven Basart, Sanmi Koyejo, Dawn Song, Matt Fredrikson, J. Zico Kolter, and Dan Hendry… — Representation Engineering: A Top-Down Approach to AI Transparency

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/scalena-2026-steering-large-language-models-machine/