A Comprehensive Evaluation of Multilingual Chain-of-Thought Reasoning: Performance, Consistency, and Faithfulness Across Languages
This work addresses the underexplored problem of multilingual reasoning traces for researchers and practitioners in natural language processing, though it is incremental as it extends existing evaluation techniques to new dimensions.
The study evaluated multilingual Chain-of-Thought reasoning in large reasoning models, finding that performance, consistency, and faithfulness of thinking traces vary significantly across languages, with models showing strong language preferences and divergent reliance on traces.
Large reasoning models (LRMs) increasingly rely on step-by-step Chain-of-Thought (CoT) reasoning to improve task performance, particularly in high-resource languages such as English. While recent work has examined final-answer accuracy in multilingual settings, the thinking traces themselves, i.e., the intermediate steps that lead to the final answer, remain underexplored. In this paper, we present the first comprehensive study of multilingual CoT reasoning, evaluating three key dimensions: performance, consistency, and faithfulness. We begin by measuring language compliance, answer accuracy, and answer consistency when LRMs are explicitly instructed or prompt-hacked to think in a target language, revealing strong language preferences and divergent performance across languages. Next, we assess crosslingual consistency of thinking traces by interchanging them between languages. We find that the quality and effectiveness of thinking traces vary substantially depending on the prompt language. Finally, we adapt perturbation-based techniques -- i.e., truncation and error injection -- to probe the faithfulness of thinking traces across languages, showing that models rely on traces to varying degrees. We release our code and data to support future research.