Building Multilingual Bridges: Data Mixing as the Pillar of Generalization for In-Language Reasoning

  • 类型:arxiv
  • 标识:2609.10445
  • 链接:https://arxiv.org/abs/2609.10445
  • 主分类:engineering
  • 形态:method
  • TLDR:Reasoning language models have made substantial advances on a variety of complex tasks, yet their capabilities remain overwhelmingly English-centric: models primarily reason in English regardless of the language they are prompted in. 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. 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.
  • 待LLM分类:是
  • 标题中文:构建多语言桥梁:数据混合作为语言内推理泛化的支柱
  • TLDR中文:推理语言模型在多种复杂任务上取得显著进展,但其能力仍以英语为中心:模型无论被提示使用何种语言,主要以英语进行推理。这对非英语用户不可及,存在丢失原问题意图的风险,也放弃了目标语言中更易表达的知识。本工作推进 L2 推理,即模型在用户提示语言中一致推理的能力,从而在提示与回答之间架起语言内的桥梁。
  • 来源文件
  • /inbox/tom/_candidates/2026-09-12-agent-rag-longcontext-candidates.json