English(EN) How Well Does Mathematical Reasoning Transfer Across Languages in English-Centric LLMs?

研究：LLM跨语言数学推理迁移能力参差不齐

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-23 04:00

一项新研究调查了主要使用英文数据训练的大型语言模型（LLM）的跨语言数学推理能力。研究发现，数学推理能力向其他语言的迁移能力差异很大，具体取决于特定的LLM、目标语言和训练方法。有趣的是，英文数学推理能力更强的模型并不总是在迁移到其他语言时表现出更有效的迁移。研究还强调，在训练中加入哪怕一种平行语言都能显著提升性能，而增加更多语言则能带来较小但持续的提升。 AI

影响强调了对LLM进行直接多语言评估的必要性，以了解其真实的跨语言推理能力。

排序理由 arXiv上发表的研究论文，详细介绍了LLM能力的研究结果。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 5 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; arXiv上发表的研究论文，详细介绍了LLM能力的研究结果。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CL TIER_1 English(EN) · Wen Yang, Junhong Wu, Chong Li, Chengqing Zong, Jiajun Zhang · 2026-10-23 04:00

How Well Does Mathematical Reasoning Transfer Across Languages in English-Centric LLMs?

arXiv:2510.02272v2 Announce Type: replace Abstract: Recent advances in Reinforcement Post-Training (RPT) have substantially improved the mathematical reasoning capabilities of large language models (LLMs), but it remains unclear how well such gains transfer across languages. In t…

本文如何被排名

报道来源 [1]

How Well Does Mathematical Reasoning Transfer Across Languages in English-Centric LLMs?

相关实体

相关话题