Work

A Comprehensive Evaluation of Multilingual Chain-of-Thought Reasoning: Performance, Consistency, and Faithfulness Across Languages

Year: 2026

Type: conference-paper

Abstract: 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 finalanswer 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 a 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.1 Method Model ar bn de en es fr hi id it ja ko pt sw yo zh Explicit Instruction

Source: Findings of the Association for Computational Linguistics: EACL

DOI: 10.18653/v1/2026.findings-eacl.276

Authors: Raoyuan Zhao, Yihong Liu, Hinrich Schuetze, Michael A. Hedderich

Institution: Ludwig-Maximilians-Universität München

Language: English

Topic: Topic Modeling

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Funders: DeepMind, Deutsche Forschungsgemeinschaft

Award: G7578079922