ReasonXL: Shifting LLM Reasoning Language Without Sacrificing Performance

• URL: http://arxiv.org/abs/2604.12378v1
• Date: Tue, 14 Apr 2026 07:10:46 GMT
• Title: ReasonXL: Shifting LLM Reasoning Language Without Sacrificing Performance
• Abstract summary: We introduce ReasonXL, the first large-scale parallel corpus of cross-domain reasoning traces spanning five European languages.<n>Using ReasonXL, we demonstrate that LLMs can be adapted to reason entirely in a desired target language.<n>We conduct an extensive representational analysis of the adaptation and find a clear functional division across model depth.
• Score: 44.67427662307542
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Despite advances in multilingual capabilities, most large language models (LLMs) remain English-centric in their training and, crucially, in their production of reasoning traces. Even when tasked with non-English problems, these models predominantly reason in English, creating a fundamental mismatch for non-English usage scenarios. We address this disparity directly with three contributions. (i) We introduce ReasonXL, the first large-scale parallel corpus of cross-domain reasoning traces spanning five European languages (English, German, French, Italian, and Spanish), with over two million aligned samples per language, each comprising prompts, reasoning traces, and final outputs, enabling direct supervision of language-specific reasoning. (ii) Using ReasonXL, we demonstrate that LLMs can be adapted to reason entirely in a desired target language, using a simple two-stage pipeline of supervised fine-tuning (SFT) followed by reinforcement learning with verifiable rewards (RLVR). The resulting models match or exceed baseline performance, with minimal loss in general knowledge and broadly preserved cross-lingual transfer. (iii) We conduct an extensive representational analysis of the adaptation and find a clear functional division across model depth: early layers contain an activation bottleneck that causally determines language identity, while upper layers concentrate the weight and activation changes driven by adaptation. We further find that RLVR achieves greater behavioral divergence from the base model with smaller parameter updates than SFT, suggesting a more efficient representational rerouting despite much smaller weight updates.

Related papers

• Building Multilingual Bridges: Data Mixing as the Pillar of Generalization for In-Language Reasoning [75.46846056993418]
  Reasoning language models primarily reason in English regardless of the language they are prompted in.<n>This is inaccessible for non-English-speaking users, risks losing the intent of the original question, and forgoes knowledge more readily expressed in the target language.<n>In this work, we advance L2 reasoning, the ability of a model to reason consistently in the language of the user's prompt, thus building an in-language bridge between the prompt and the answer.
  arXiv  Detail & Related papers  (2026-09-09T16:56:25Z)
• Choosing the Right Language Mode at Inference Time for Multilingual Reliability [3.6095388702618405]
  Large language models often struggle to reason in low- to mid-resource languages.<n>Prior work has shown that translation can improve multilingual reasoning by helping models access stronger English-centric representations.<n>This raises a central question: How much translation is needed for multilingual large language models to reason reliably.
  arXiv  Detail & Related papers  (2026-09-04T02:36:26Z)
• Efficient Multilingual Reasoning Transfer via Progressive Code-Switching [51.49388869532955]
  Transfer approaches typically rely on distilled target-language reasoning traces from stronger LRMs or online supervision from external judge models.<n>We propose PCS (Progressive Code-Switching), a more efficient transfer framework that requires only lightweight translation.<n> PCS substantially narrows the performance gap between target-language and English reasoning, yielding more language-consistent reasoning.
  arXiv  Detail & Related papers  (2026-07-01T06:13:16Z)
• AdaMame: A Training Recipe for Adaptive Multilingual Reasoning [32.31269959716733]
  Large Reasoning Models (LRMs) often fail to reason in the language of the query.<n>AdaMame is a two-stage training recipe for multilingual mathematical reasoning.
  arXiv  Detail & Related papers  (2026-06-13T03:22:35Z)
• Align to the Pivot: Dual Alignment with Self-Feedback for Multilingual Math Reasoning [71.4175109189942]
  We present Pivot-Aligned Self-Feedback Multilingual Reasoning (PASMR)<n>This approach designates the model's primary language as the pivot language.<n>It establishes a cross-lingual self-feedback mechanism without relying on external correct answers or reward models.
  arXiv  Detail & Related papers  (2026-01-25T03:20:00Z)
• Code-Switching In-Context Learning for Cross-Lingual Transfer of Large Language Models [64.54005959758733]
  We introduce code-switching in-context learning (CSICL) as a principled and robust approach for overcoming the translation barrier during inference.<n>We conduct extensive experiments across 4 LLMs, 6 datasets, and 10 languages, spanning both knowledge-intensive and reasoning-oriented domains.<n>Our results demonstrate CSICL consistently outperforms X-ICL baselines, achieving gains of 3.1%p and 1.9%p in both target and unseen languages.
  arXiv  Detail & Related papers  (2025-10-07T08:35:42Z)
• Parallel Scaling Law: Unveiling Reasoning Generalization through A Cross-Linguistic Perspective [52.452449102961225]
  This study proposes a novel cross-linguistic perspective to investigate reasoning generalization.<n>Our findings reveal that cross-lingual transferability varies significantly across initial model, target language, and training paradigm.<n>Our study challenges the assumption that LRM reasoning mirrors human cognition, providing critical insights for the development of more language-agnostic LRMs.
  arXiv  Detail & Related papers  (2025-10-02T17:49:49Z)
• Cross-lingual Collapse: How Language-Centric Foundation Models Shape Reasoning in Large Language Models [44.94287386776289]
  We identify textbfCross-lingual Collapse, a systematic drift in which a multilingual language model reverts to its dominant pre-training language.<n>Our experiments reveal three key findings: (i) GRPO rapidly amplifies pre-training language imbalances, leading to the erosion of low-resource languages within just a few hundred updates; (ii) language consistency reward mitigates this drift but does so at the expense of an almost 5 - 10 pp drop in accuracy.
  arXiv  Detail & Related papers  (2025-06-06T08:08:48Z)
• The Unreasonable Effectiveness of Model Merging for Cross-Lingual Transfer in LLMs [45.08958917457921]
  Large language models (LLMs) still struggle across tasks outside of high-resource languages.<n>In this work, we investigate cross-lingual transfer to lower-resource languages where task-specific post-training data is scarce.
  arXiv  Detail & Related papers  (2025-05-23T20:28:31Z)
• When Less Language is More: Language-Reasoning Disentanglement Makes LLMs Better Multilingual Reasoners [111.50503126693444]
  We show that language-specific ablation consistently boosts multilingual reasoning performance.<n>Compared to post-training, our training-free ablation achieves comparable or superior results with minimal computational overhead.
  arXiv  Detail & Related papers  (2025-05-21T08:35:05Z)
• Enhancing Non-English Capabilities of English-Centric Large Language Models through Deep Supervision Fine-Tuning [42.166438218926274]
  We introduce a deep supervision fine-tuning method (DFT) that incorporates additional supervision in the internal layers of the model to guide its workflow.<n>Our method guides the model to not only consider the final generated result when processing non-English inputs but also ensure the accuracy of internal representations.
  arXiv  Detail & Related papers  (2025-03-03T07:59:32Z)
• Eliciting Better Multilingual Structured Reasoning from LLMs through Code [17.870002864331322]
  We introduce a multilingual structured reasoning and explanation dataset, termed xSTREET, that covers four tasks across six languages. xSTREET exposes a gap in base LLM performance between English and non-English reasoning tasks. We propose two methods to remedy this gap, building on the insight that LLMs trained on code are better reasoners.
  arXiv  Detail & Related papers  (2024-03-05T00:48:56Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.