Language of Thought Shapes Output Diversity in Large Language Models

TL;DR

Controlling the language of thought in large language models significantly enhances output diversity, with multilingual thinking strategies expanding the model's creative and cultural coverage.

Contribution

This work introduces the concept of the language of thought as a structural source of diversity and demonstrates its effectiveness through extensive multilingual sampling experiments.

Findings

Switching to non-English thinking languages increases output diversity.

Languages farther from English in the thinking space yield larger diversity gains.

Aggregating multiple thinking languages further enhances diversity.

Abstract

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.