ReasonXL: Shifting LLM Reasoning Language Without Sacrificing Performance

Researchers introduce ReasonXL, a large-scale parallel corpus of reasoning traces in five European languages, enabling LLMs to reason entirely in non-English languages through a supervised fine-tuning and reinforcement learning pipeline. This adaptation allows models to maintain high performance while better serving non-English speaking users, with minimal loss in general knowledge and preserved cross-lingual transfer capabilities.

Read the full article at arXiv cs.CL (NLP)

---

Want to create content about this topic? Use Nemati AI tools to generate articles, social posts, and more.