arXiv：cs.CL· Oliver Hauck, Mario Sanz-Guerrero, Katharina von der Wense·· 4 小时前AI 评分37
研究推理语言对齐在单语 RAG 中的作用
Investigating the Role of Reasoning-Language Alignment in Monolingual Retrieval-Augmented Generation
AI 导读
研究者构建了完全单语的德语 RAG 问答测试平台，基于桌面角色扮演游戏《黑眼》的虚构世界，该领域德语资料丰富但过于小众，模型无法凭记忆作答，必须依赖检索。实验发现，强制德语推理优于强制法语，尽管模型在法语上基准分数更高，说明收益来自语言对齐而非语言能力；检索上下文越丰富、结构感知越强，优势越明显。但强制德语仅达到模型原生无约束英语推理的水平，未能超越，表明原生多语言推理仍然必要。
正文
Abstract:Reasoning traces improve large language models (LLMs), but current models are trained to reason mostly in English. It has been shown that forcing a model to reason in another language degrades accuracy, even when the reasoning language matches the language of the prompt -- but only for a setting where the model reasons over a short prompt. Here, we ask whether the same holds for retrieval-augmented generation (RAG), where the model must read and integrate a large amount of retrieved evidence in the target language. To study this, we build a fully monolingual German RAG question-answering testbed over the fictional world of the tabletop role-playing game The Dark Eye, a domain that is richly documented in German but too niche for the model to answer from memory, so that it has to rely on retrieval. Varying the forced reasoning language of an agentic RAG system on this testbed, we find that aligning the reasoning language with the language of the query and the retrieved documents helps. Forced German reasoning outperforms forced French, although the model benchmarks higher in French, so the benefit comes from alignment and not from language proficiency. The advantage grows when the retrieved context is richer and structure-aware. However, forced German only reaches the level of the model's native, unconstrained English reasoning without surpassing it, showing that native multilingual reasoning is needed. We publicly release the testbed and QA benchmark.
| Comments: | Accepted to the Workshop on Open Reasoning Across Cultures & Languages at EMNLP 2026 |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.03136 [cs.CL] |
| (or arXiv:2610.03136v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03136 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Mario Sanz-Guerrero [view email]
[v1]
Fri, 2 Oct 2026 11:02:04 UTC (530 KB)
来源：arXiv：cs.CL · arxiv.org