Logo Lanfrica
  • Accueil
  • Atlas
  • Analyses
  • Documentation
  • Sign in

© 2026 Lanfrica. Tous droits réservés. Tous les droits d'auteur des ressources affichées sur le site Web Lanfrica appartiennent aux détenteurs de droits d'auteur d'origine, sauf indication contraire explicite.

Unlocking Multilingual Reasoning Capability of LLMs and LVLMs through Representation Engineering

Domaine:

natural language processing

Type de record:

paper
Créateur:
Li,FenMa,Ye,
Hôte:avatar
Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) demonstrate strong reasoning capabilities, yet their performance in English significantly outperforms that in low-resource languages, raising fairness concerns in multilingual applications. Existing approaches either rely on costly multilingual training or employ prompting with external translation tools, both of which are resource-intensive and sensitive to translation quality. To address these limitations, we propose a training-free inference-time method to enhance Multilingual Reasoning capabilities via Representation Engineering (MRRE) without using any additional training data or tools. MRRE sequentially injects two precomputed vectors at specific layers during inference processing: cross-lingual reasoning enhancement vectors, which steer non-English reasoning representations toward English space to unlock multilingual reasoning, and target-language output anchoring vectors, which restore the distribution of the target language to preserve input-output language consistency. Comprehensive experiments across six advanced LLMs and LVLMs on four reasoning benchmarks demonstrate that MRRE consistently enhances non-English reasoning by an average gain of 5.48% and up to 7.54% in low-resource languages (Thai and Swahili), while improving input-output language consistency by 3.78%.

Visit

arxiv.org

Tasks

commonsense reasoning

Languages

Swahili

Tags

Computer Vision and Pattern Recognition

Similaires

EMCEE: Improving Multilingual Capability of LLMs via Bridging Knowledge and Reasoning with Extracted Synthetic Multilingual ContextDo Moral Judgment and Reasoning Capability of LLMs Change with Language? A Study using the Multilingual Defining Issues TestLLMs Beyond English: Scaling the Multilingual Capability of LLMs with Cross-Lingual FeedbackOptimal Transport Distillation for Zero-Shot Cross-Lingual Reasoning in Multilingual LLMsFine-tuning LLMs on Intermediate Text Classification for Multilingual Zero-shot ReasoningAnalyzing LLMs' Cognition of Knowledge Boundary Across Languages Through the Lens of Internal Representation

EMCEE: Improving Multilingual Capability of LLMs via Bridging Knowledge and Reasoning with Extracted Synthetic Multilingual Context

Large Language Models (LLMs) have achieved impressive progress across a wide range of tasks, yet the

Do Moral Judgment and Reasoning Capability of LLMs Change with Language? A Study using the Multilingual Defining Issues Test

This paper explores the moral judgment and moral reasoning abilities exhibited by Large Language Mod

LLMs Beyond English: Scaling the Multilingual Capability of LLMs with Cross-Lingual Feedback

To democratize large language models (LLMs) to most natural languages, it is imperative to make thes

Optimal Transport Distillation for Zero-Shot Cross-Lingual Reasoning in Multilingual LLMs

Benefiting from transformer-based pre-trained language models, neural ranking models have made signi

Fine-tuning LLMs on Intermediate Text Classification for Multilingual Zero-shot Reasoning

Intermediate-task training---fine-tuning a pretrained model on an intermediate task before fine-tuni

Analyzing LLMs' Cognition of Knowledge Boundary Across Languages Through the Lens of Internal Representation

While understanding the knowledge boundaries of LLMs is crucial to prevent hallucination, research o