Cross-Lingual Exploration for Parametric Knowledge

2026-06-23

AI Summary

This work addresses the uneven accessibility of parametric knowledge across languages in large language models, which hinders cross-lingual knowledge transfer and consistency. To mitigate this limitation without model expansion, the authors propose a cross-lingual prompting strategy that activates latent multilingual factual knowledge during inference. The approach systematically explores the multilingual prompt space along four intrinsic dimensions and is validated through cross-lingual prompt engineering, evaluation on a multilingual factual benchmark, and analysis of parametric knowledge retrieval. Experiments across 17 typologically diverse languages demonstrate that the proposed method significantly improves factual recall and cross-lingual consistency while achieving higher computational efficiency compared to baseline approaches relying on native-language model extensions.

Abstract

Parametric knowledge in Large Language Models is not equally accessible across languages. As a result, standard inference techniques often struggle to surface localized facts, leading to failures in cross-lingual knowledge transfer and consistency. In this work, we investigate techniques for accessing hidden factual knowledge by exploring cross-lingual prompting strategies. We identify four inherent dimensions of cross-lingual exploration that directly govern parametric knowledge retrieval and evaluate them on multilingual factual benchmarks covering 17 typologically diverse languages. Our results demonstrate that cross-lingual exploration significantly improves knowledge transfer and factual recall, representing a more efficient compute Pareto frontier than native-language scaling. Furthermore, we observe corresponding improvements in cross-lingual consistency, exceeding what can be explained by accuracy gains alone. Overall, our work establishes multilingual prompt exploration as a highly effective inference-time strategy for unlocking latent parametric knowledge.

Problem

• cross-lingual
• parametric knowledge
• knowledge transfer
• factual recall
• multilingual

Innovation

• cross-lingual prompting
• parametric knowledge
• multilingual inference
• knowledge retrieval
• language model consistency