SOURCE-LINKED INTELLIGENCE

Choosing the Right Language Mode at Inference Time for Multilingual Reliability

arXiv · AI, language, vision and robotics · article · Sep 4, 2026 · UTC

Multilingual large language models often struggle to reason in low- to mid-resource languages. Prior work has shown that translation can improve multilingual reasoning by helping models access stronger English-centric representations. This raises a central question: How much translation is needed for multilingual large language models to reason reliably, and when does more translation instead trigger interference and overconfidence? Using LLaMA and Qwen models, we run extensive experiments varying text scope and language mode (target-only, English-only, bilingual) to evaluate both accuracy and

Read original source ↗ Open in workspace

recordType

paper

region

Global