Elamur Research⌕
☰
home / papers / 2026 / Let's Simplify Step by Step: Guiding LLM Towards Multilingual Unsupervised Proficiency-Controlled Sentence Simplification

Let's Simplify Step by Step: Guiding LLM Towards Multilingual Unsupervised Proficiency-Controlled Sentence Simplification

Summary

LLMs are weak at rewriting a sentence to a specific CEFR level when the gap between source and target is large, for example from C2 down to A1. The authors split such large jumps into a chain of smaller ones. They first run simplifications on the validation data and check the results with a sentence-level CEFR estimator. From these checks they build a reward matrix that scores how reliably the model moves between each pair of levels. Dynamic programming then finds the best sequence of intermediate levels. For each step in that sequence, the prompt receives up to three validated exemplar pairs, chosen for the highest semantic similarity (STS) between original and output. Generation runs as chain-of-thought inside a single conversation, so the model sees its earlier steps. Experiments used Llama and Gemma on CEFR-SP and README++ across five languages. Target-level accuracy improved and fewer inference steps were needed. Human evaluation showed that stronger simplification costs meaning, and annotators themselves struggled to agree on meaning preservation. Faithful large-span simplification therefore remains open.

Key points

  • Languages and data: English, French, Russian, Arabic and Hindi, evaluated on CEFR-SP and README++.
    • The method needs no supervised training. It works through prompting, guided by the validation data.
  • Path planning: level transitions are scored +1 for an exact target match, +0.5 for a miss by one level, and −1 otherwise. Scores are min-max normalized, then dynamic programming picks the path.
    • Example: C2→A1 is routed through B2, B1 and A2.
  • Reported gains:
    • Target-level achievement improves by up to 20 percentage points over one-step prompting.
    • Inference steps drop by 22–42% compared with simplifying through every level in sequence, while quality stays similar. On Llama the reduction is 25% for most corpora and 22.2% for README_FR.
  • Human readability check (Planning + Semantic Chat History):
    • Exact-level accuracy is about 53–62% and adjacent-level accuracy 95–100%.
    • Inter-annotator Krippendorff's Alpha ranges from 0.62 to 0.79.
  • Meaning-preservation annotation: native speakers agreed less than language instructors did. For French and Hindi, only instructor judgments were used.
    • Code is released at github.com/JasonZhang0702/simplify-step-by-step.

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

Fields: Machine translation · Natural language processing

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-eacl.279

@inproceedings{zhang2026let,
  title = {{Let's Simplify Step by Step: Guiding LLM Towards Multilingual Unsupervised Proficiency-Controlled Sentence Simplification}},
  author = {Zhang, Jingshen and Qiu, Xin Ying and Lu, Lifang and Huang, Zhuhua and Hu, Yutao and Wu, Yuechang and Lu, JunYu},
  booktitle = {Findings of the Association for Computational Linguistics: EACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-eacl.279},
  url = {https://doi.org/10.18653/v1/2026.findings-eacl.279}
}

References

44 references, 12 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Sweta Agrawal and Marine Carpuat. 2023. Controlling pre-trained language models for grade-specific text simplification. In Proceedings of ACL 2023. ↗
  2. Kiyomi Chujo, Kathryn Oghigian, and Shiro Akasegawa. 2015. A corpus and grammatical browsing system for remedial efl learners. In Agnieszka Lenko-Szymańska and Alex Boulton, editors, Multiple Affordances of Language Corpora for Data-driven Learning, volume 69 of Studies in Corpus Linguistics, pages 109–128. John Benjamins. ↗
  3. Liam Cripwell, Joël Legrand, and Claire Gardent. 2023. Document-level planning for text simplification. In Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics, pages 993–1006, Dubrovnik, Croatia. Association for Computational Linguistics. find
  4. Fernando Alva-Manchego, Joachim Bingel, Gustavo Paetzold, Carolina Scarton, and Lucia Specia. 2017. Learning how to simplify from explicit labeling of complex-simplified text pairs. In Proceedings of the Eighth International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 295–305, Taipei, Taiwan. Asian Federation of Natural Language Processing. find our card
  5. Yue Dong, Zichao Li, Mehdi Rezagholizadeh, and Jackie Chi Kit Cheung. 2019. Editnts: A neural programmer-interpreter model for sentence simplification through explicit editing. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (ACL), pages 3393–3402, Florence, Italy. Association for Computational Linguistics. find
  6. Yuki Arase, Satoru Uchida, and Tomoyuki Kajiwara. 2022. Cefr-based sentence difficulty annotation and assessment. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 6206–6219. find
  7. Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, et al. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783. ↗
  8. Abdulelah Barayan, Jose Camacho-Collados, and Fernando Alva-Manchego. 2025. Analysing zero-shot readability-controlled sentence simplification. In Proceedings of the 31st International Conference on Computational Linguistics, pages 6762–6781. ↗
  9. Arian Farajidizaji, Vinit Raina, and Mark Gales. 2024. Is it possible to modify text to a target readability level? an initial investigation using zero-shot large language models. arXiv preprint arXiv:2309.12551. ↗
  10. Sofia Blinova, Xinyu Zhou, Martin Jaggi, Carsten Eickhoff, and Seyed Ali Bahrainian. 2023. SIMSUM: Document-level text simplification via simultaneous summarization. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 9927– 9944, Toronto, Canada. Association for Computational Linguistics. find
  11. Goran Glavaš and Sanja Štajner. 2015. Simplifying lexical simplification: Do we need simplified corpora? In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 2: Short Papers), pages 63–68, Beijing, China. Association for Computational Linguistics. find our card
  12. Natalia Grabar and Rémi Cardon. 2018. CLEAR – simple corpus for medical French. In Proceedings of the 1st Workshop on Automatic Text Adaptation (ATA), pages 3–9, Tilburg, the Netherlands. Association for Computational Linguistics. find
  13. Joseph Marvin Imperial, Abdullah Barayan, Regina Stodden, Rodrigo Wilkens, Ricardo Munoz Sanchez, Lingyun Gao, Melissa Torgbi, Dawn Knight, Gail Forey, Reka R. Jablonkai, Ekaterina Kochmar, Robert Reynolds, Eugénio Ribeiro, Horacio Saggion, Elena Volodina, Sowmya Vajjala, Thomas François, Fernando Alva-Manchego, and Harish Tayyar Madabushi. 2025. Universalcefr: Enabling open multilingual research on language proficiency assessment. Preprint, arXiv:2506.01419. ↗
  14. Joseph Marvin Imperial and Harish Tayyar Madabushi. 2023. Flesch or fumble? evaluating readability standard alignment of instruction-tuned language models. In Proceedings of the Third Workshop on Natural Language Generation, Evaluation, and Metrics (GEM), pages 205–223. ↗
  15. L Javourey-Drevet, S Dufau, T François, N Gala, J Ginestié, and JC Ziegler. 2022. Simplification of literary and scientific texts to improve reading fluency and comprehension in beginning readers of French. Applied Psycholinguistics, 43(2):485–512. ↗
  16. Albert Qiaochu Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de Las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, L’elio Renard Lavaud, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas Wang, Timothée Lacroix, and William El Sayed. 2023. Mistral 7b. arXiv preprint arXiv:2310.06825. ↗
  17. Chao Jiang, Mounica Maddela, Wuwei Lan, Yang Zhong, and Wei Xu. 2020. Neural crf model for sentence alignment in text simplification. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 7943– 7960, Online. Association for Computational Linguistics. find our card
  18. Diederik P. Kingma and Jimmy Ba. 2017. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980. ↗
  19. Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, and Yusuke Iwasawa. 2022. Large language models are zero-shot reasoners. In Advances in Neural Information Processing Systems (NeurIPS), volume 35, pages 22199–22213. Curran Associates, Inc. find our card
  20. Jiachang Liu, Dinghan Shen, Yizhe Zhang, William B. Dolan, Lawrence Carin, and Weizhu Chen. 2022. What makes good in-context examples for gpt-3? In Proceedings of Deep Learning Inside Out (Dee-LIO 2022): The 3rd Workshop on Knowledge Extraction and Integration for Deep Learning Architectures, pages 100–114, Dublin, Ireland. Association for Computational Linguistics. ↗ our card
  21. Mounica Maddela, Fernando Alva-Manchego, and Wei Xu. 2021. Controllable text simplification with explicit paraphrasing. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 3536– 3553, Online. Association for Computational Linguistics. find our card
  22. Ali Malik, Stephen Mayhew, Chris Piech, and Klinton Bicknell. 2024. From tarzan to tolkien: Controlling the language proficiency level of llms for content generation. Preprint, arXiv:2406.03030. ↗
  23. Louis Martin, Éric Villemonte de la Clergerie, Benoît Sagot, and Antoine Bordes. 2020. Controllable sentence simplification. In Proceedings of the Twelfth Language Resources and Evaluation Conference (LREC), pages 4689–4698, Marseille, France. European Language Resources Association. ↗
  24. Matej Martinc, Senja Pollak, and Marko Robnik-Šikonja. 2021. Supervised and unsupervised neural approaches to text readability. Computational Linguistics, 47(1):141–179. ↗
  25. Kaijie Mo and Renfen Hu. 2024. ExpertEase: A multi-agent framework for grade-specific document simplification with large language models. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 9080–9099, Miami, Florida, USA. Association for Computational Linguistics. ↗
  26. Amit Moryossef, Yoav Goldberg, and Ido Dagan. 2019. Step-by-step: Separating planning from realization in neural data-to-text generation. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 2267–2277, Minneapolis, Minnesota, USA. Association for Computational Linguistics. ↗ our card
  27. Tarek Naous, Michael J Ryan, Anton Lavrouk, Mukesh Chandra, and Wei Xu. 2024. Readme++: Benchmarking multilingual language models for multi-domain readability assessment. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 12230–12266. find
  28. Masashi Oshika, Makoto Morishita, Tsutomu Hirao, Ryohei Sasano, and Koichi Takeda. 2024. Simplifying translations for children: Iterative simplification considering age of acquisition with llms. Preprint, arXiv:2408.04217. ↗
  29. Gustavo Paetzold and Lucia Specia. 2015. Lexenstein: A framework for lexical simplification. In Proceedings of ACL-IJCNLP 2015 System Demonstrations, pages 85–90, Beijing, China. Association for Computational Linguistics. find
  30. Ratish Puduppully, Li Dong, and Mirella Lapata. 2019. Data-to-text generation with content selection and planning. In Proceedings of the AAAI Conference on Teven Le Scao, Sylvain Gugger, Mariama Drame, Quentin Lhoest, and Alexander Rush. 2020. Transformers: State-of-the-art natural language processing. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 38–45, Online. Association for Computational Linguistics. Artificial Intelligence, volume 33, pages 6908–6915, Honolulu, Hawaii, USA. AAAI Press. ↗ our card
  31. Jipeng Qiang, Yun Li, Yi Zhu, Yunhao Yuan, Yang Shi, and Xindong Wu. 2021. Lsbert: Lexical simplification based on bert. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 29:3064– 3076. ↗
  32. Nils Reimers and Iryna Gurevych. 2019. Sentence-bert: Sentence embeddings using siamese bert-networks. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 3982–3992, Hong Kong, China. Association for Computational Linguistics. ↗
  33. Michael Ryan, Tarek Naous, and Wei Xu. 2023. Revisiting non-English text simplification: A unified multilingual benchmark. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 4898– 4927, Toronto, Canada. Association for Computational Linguistics. find
  34. Andrey Sakhovskiy, Alexandra Izhevskaya, Alena Pestova, Elena Tutubalina, Valentin Malykh, Ivana Smurov, and Ekaterina Artemova. 2021. RuSimpleSentEval-2021 shared task: Evaluating sentence simplification for russian. In Proceedings of the International Conference “Dialogue”, pages 607–617. find
  35. Carolina Scarton and Lucia Specia. 2018. Learning simplifications for specific target audiences. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 712–718. find
  36. Kevin Stowe, Debanjan Ghosh, and Mengxuan Zhao. 2022. Controlled language generation for language learning items. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing: Industry Track (EMNLP Industry Track), pages 294–305, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics. find
  37. Gemma Team, Morgane Riviere, Shreya Pathak, Pier Giuseppe Sessa, Cassidy Hardin, Surya Bhupatiraju, Léonard Hussenot, et al. 2024. Gemma 2: Improving open language models at a practical size. arXiv preprint arXiv:2408.00118. ↗
  38. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed Chi, Quoc Le, and Denny Zhou. 2023. Chain-of-thought prompting elicits reasoning in large language models. arXiv preprint arXiv:2201.11903. ↗ our card
  39. Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pier-ric Cistac, Tim Rault, Remi Louf, Morgan Funtowicz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Wei Xu, Chris Callison-Burch, and Courtney Napoles. 2015. Problems in current text simplification research: New data can help. Transactions of the Association for Computational Linguistics, 3:283–297. find our card
  40. Wei Xu, Courtney Napoles, Ellie Pavlick, Quanze Chen, and Chris Callison-Burch. 2016. Optimizing statistical machine translation for text simplification. Transactions of the Association for Computational Linguistics, 4:401–415. ↗ our card
  41. Bohan Zhang, Prafulla Kumar Choubey, and Ruihong Huang. 2022. Predicting sentence deletions for text simplification using a functional discourse structure. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 255–261, Dublin, Ireland. Association for Computational Linguistics. find
  42. Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q. Weinberger, and Yoav Artzi. 2020. Bertscore: Evaluating text generation with bert. In International Conference on Learning Representations. ↗
  43. Yang Zhong, Chao Jiang, Wei Xu, and Junyi Jessy Li. 2020. Discourse level factors for sentence deletion in text simplification. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pages 9709–9716, New York, New York, USA. AAAI Press. find
  44. Denny Zhou, Nathanael Schärli, Le Hou, Jason Wei, Nathan Scales, Xuezhi Wang, Dale Schuurmans, et al. 2023. Least-to-most prompting enables complex reasoning in large language models. In Proceedings of the Eleventh International Conference on Learning Representations (ICLR). OpenReview.net. ↗ our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/zhang-2026-let-s-simplify-step-step/