连续思维比显式推理更适配多语言，尤其在低资源语言上表现更好。
Is continuous CoT better suited for multi-lingual reasoning?
- 用连续潜空间进行思维推理（CODI框架）
- 在低资源语言零样本场景下性能提升29至50倍压缩效率
- 适合需要跨语言推理的高效模型部署
我们探究在连续潜在空间中进行推理是否能提升多语言能力。对比了连续思维链（使用CODI框架）与标准监督微调，在五种语言类型差异较大的语言（英语、中文、德语、法语、乌尔都语）上进行实验。在GSM8k和CommonsenseQA数据集上的结果表明，连续推理在低资源语言上显著优于显式推理，尤其在目标语言未在训练中出现的零样本设置下。此外，该方法实现极高效率，推理轨迹压缩约29至50倍。这些发现表明，连续潜在表示天然具备更强的语言不变性，为跨语言推理提供可扩展的解决方案。
原文摘要 · Abstract (English)
We investigate whether performing reasoning in a continuous latent space leads to more robust multilingual capabilities. We compare Continuous Chain-of-Thought (using the CODI framework) against standard supervised fine-tuning across five typologically diverse languages: English, Chinese, German, French, and Urdu. Our experiments on GSM8k and CommonsenseQA demonstrate that continuous reasoning significantly outperforms explicit reasoning on low-resource languages, particularly in zero-shot settings where the target language was not seen during training. Additionally, this approach achieves extreme efficiency, compressing reasoning traces by approximately $29\times$ to $50\times$. These findings indicate that continuous latent representations naturally exhibit greater language invariance, offering a scalable solution for cross-lingual reasoning.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。