Language of Thought Shapes Output Diversity in Large Language Models
This addresses the need for pluralism and creativity in LLM outputs, offering a novel structural approach to enhance diversity, though it is incremental as it builds on existing multilingual capabilities.
The study tackled the problem of output diversity in Large Language Models by controlling the language used during model thinking, revealing that switching from English to non-English thinking languages consistently increases diversity, with gains up to 30% for languages farther from English, and aggregating across multiple languages yields further improvements.
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.