语言形态影响大语言模型的输出多样性
Language of Thought Shapes Output Diversity in Large Language Models
AI总结：
本文研究了语言形态对大语言模型输出多样性的影响，通过多语言思考策略提升输出多样性，并在实际应用中展现文化知识的广泛覆盖。
摘要：
输出多样性对于大语言模型至关重要，因为它支撑了多元主义和创造力。在本工作中，我们揭示了在模型思考过程中控制使用的语言——即思维语言——提供了一个新的且结构性的输出多样性来源。我们的初步研究显示，不同的思维语言占据模型思考空间中的不同区域。基于这一观察，我们研究了两种重复采样策略——单语言采样和混合语言采样——并在控制输出为英语的情况下进行多样性评估，无论使用的思维语言如何。在广泛的实验中，我们证明了将思维语言从英语切换到非英语语言会一致地增加输出多样性，呈现出清晰且一致的正相关关系，即在思维空间中距离英语越远的语言产生的增益越大。我们进一步展示，通过跨多种思维语言聚合样本可以获得额外的改进，通过组合效应实现，且通过增加语言异质性的采样规模可以扩大模型的多样性上限。最后，我们证明这些发现可以转化为多元对齐场景中的实际效益，导致LLM输出中文化知识和价值取向的更广泛覆盖。我们的代码在https://github.com/iNLP-Lab/Multilingual-LoT-Diversity上公开。
英文摘要：
Output diversity is crucial for Large Language Models as it underpins pluralism and creativity. In this work, we reveal that controlling the language used during model thinking-the language of thought-provides a novel and structural source of output diversity. Our preliminary study shows that different thinking languages occupy distinct regions in a model's thinking space. Based on this observation, we study two repeated sampling strategies under multilingual thinking-Single-Language Sampling and Mixed-Language Sampling-and conduct diversity evaluation on outputs that are controlled to be in English, regardless of the thinking language used. Across extensive experiments, we demonstrate that switching the thinking language from English to non-English languages consistently increases output diversity, with a clear and consistent positive correlation such that languages farther from English in the thinking space yield larger gains. We further show that aggregating samples across multiple thinking languages yields additional improvements through compositional effects, and that scaling sampling with linguistic heterogeneity expands the model's diversity ceiling. Finally, we show that these findings translate into practical benefits in pluralistic alignment scenarios, leading to broader coverage of cultural knowledge and value orientations in LLM outputs. Our code is publicly available at https://github.com/iNLP-Lab/Multilingual-LoT-Diversity.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
