Large Reasoning Models Are (Not Yet) Multilingual Latent Reasoners

ACL ARR 2026 January Submission2243 Authors

Keywords: multilinguality, latent reasoning, chain-of-thought

Abstract:

Large reasoning models (LRMs) achieve strong performance on mathematical reasoning tasks, often attributed to their capability to generate explicit chain-of-thought (CoT) explanations.
However, recent work shows that LRMs often arrive at the correct answer before completing these textual reasoning steps, indicating the presence of latent reasoning -- internal, non-verbal computation encoded in hidden states.
While this phenomenon has been explored in English, its multilingual behavior remains largely unknown.
In this paper, we conduct a systematic investigation of multilingual latent reasoning in LRMs across 11 languages.
Using a truncation-based strategy, we examine how the correct answer emerges as the model is given only partial reasoning traces, allowing us to measure stepwise latent prediction formation.
Our results reveal clear evidence of multilingual latent reasoning, though unevenly: strong in resource-rich languages, weaker in low-resource ones, and broadly less observable on harder benchmarks.
To understand whether these differences reflect distinct internal mechanisms, we further perform representational analyses.
Despite surface-level disparities, we find that the internal evolution of predictions is highly consistent across languages and broadly aligns with English -- a pattern suggesting an English-centered latent reasoning pathway.

Paper Type: Long

Research Area: Multilinguality and Language Diversity

Research Area Keywords: multilingualism, multilingual evaluation, chain-of-thought, reasoning

Contribution Types: Model analysis & interpretability, NLP engineering experiment, Data analysis

Languages Studied: Bengali, Chinese, French, German, Japanese, Korean, Russian, Spanish, Swahili, Telugu, Thai

Submission Number: 2243