Rethinking Cross-lingual Alignment: Balancing Transfer and Cultural Erasure in Multilingual LLMs
By: HyoJung Han, Sweta Agrawal, Eleftheria Briakou
Potential Business Impact:
Helps AI speak different languages without forgetting culture.
Cross-lingual alignment (CLA) aims to align multilingual representations, enabling Large Language Models (LLMs) to seamlessly transfer knowledge across languages. While intuitive, we hypothesize, this pursuit of representational convergence can inadvertently cause "cultural erasure", the functional loss of providing culturally-situated responses that should diverge based on the query language. In this work, we systematically analyze this trade-off by introducing a holistic evaluation framework, the transfer-localization plane, which quantifies both desirable knowledge transfer and undesirable cultural erasure. Using this framework, we re-evaluate recent CLA approaches and find that they consistently improve factual transfer at the direct cost of cultural localization across all six languages studied. Our investigation into the internal representations of these models reveals a key insight: universal factual transfer and culturally-specific knowledge are optimally steerable at different model layers. Based on this finding, we propose Surgical Steering, a novel inference-time method that disentangles these two objectives. By applying targeted activation steering to distinct layers, our approach achieves a better balance between the two competing dimensions, effectively overcoming the limitations of current alignment techniques.
Similar Papers
Bridging Language Gaps: Advances in Cross-Lingual Information Retrieval with Multilingual LLMs
Information Retrieval
Finds information in any language.
Can you map it to English? The Role of Cross-Lingual Alignment in Multilingual Performance of LLMs
Computation and Language
Helps computers understand many languages without extra training.
Analyzing and Improving Cross-lingual Knowledge Transfer for Machine Translation
Computation and Language
Helps computers translate many languages better.