Unveiling Linguistic Regions in Large Language Models

TL;DR

This paper investigates the intrinsic linguistic regions within large language models, revealing a core region vital for multilingual competence and demonstrating how perturbations affect language performance and how freezing this core can mitigate catastrophic forgetting.

Contribution

The study identifies a core linguistic region in LLMs, explores its properties, and shows how manipulating this region impacts multilingual performance and continual learning.

Findings

A core region accounts for ~1% of parameters and is crucial for linguistic competence.

Perturbing specific parameters causes significant performance drops across languages.

Freezing the core region during further training reduces catastrophic forgetting.

Abstract

Large Language Models (LLMs) have demonstrated considerable cross-lingual alignment and generalization ability. Current research primarily focuses on improving LLMs' cross-lingual generalization capabilities. However, there is still a lack of research on the intrinsic mechanisms of how LLMs achieve cross-lingual alignment. From the perspective of region partitioning, this paper conducts several investigations on the linguistic competence of LLMs. We discover a core region in LLMs that corresponds to linguistic competence, accounting for approximately 1% of the total model parameters. Removing this core region by setting parameters to zero results in a significant performance decrease across 30 different languages. Furthermore, this core region exhibits significant dimensional dependence, perturbations to even a single parameter on specific dimensions leading to a loss of linguistic competence. Moreover, we discover that distinct monolingual regions exist for different languages, and disruption to these specific regions substantially reduces the LLMs' proficiency in those corresponding languages. Our research also indicates that freezing the core linguistic region during further pre-training can mitigate the issue of catastrophic forgetting (CF), a common phenomenon observed during further pre-training of LLMs. Overall, exploring the LLMs' functional regions provides insights into the foundation of their intelligence.