The Model Arena for Cross-lingual Sentiment Analysis: A Comparative Study in the Era of Large Language Models

TL;DR

This study compares the cross-lingual sentiment analysis capabilities of small multilingual models and large language models, revealing that small models excel in zero-shot transfer, while large models adapt better in few-shot settings.

Contribution

It provides an empirical comparison of public small multilingual models and large language models for cross-lingual sentiment analysis across multiple languages.

Findings

Small multilingual models outperform LLMs in zero-shot transfer.

Large language models show better adaptation in few-shot scenarios.

Proprietary GPT models excel in zero-shot but lag in few-shot settings.

Abstract

Sentiment analysis serves as a pivotal component in Natural Language Processing (NLP). Advancements in multilingual pre-trained models such as XLM-R and mT5 have contributed to the increasing interest in cross-lingual sentiment analysis. The recent emergence in Large Language Models (LLM) has significantly advanced general NLP tasks, however, the capability of such LLMs in cross-lingual sentiment analysis has not been fully studied. This work undertakes an empirical analysis to compare the cross-lingual transfer capability of public Small Multilingual Language Models (SMLM) like XLM-R, against English-centric LLMs such as Llama-3, in the context of sentiment analysis across English, Spanish, French and Chinese. Our findings reveal that among public models, SMLMs exhibit superior zero-shot cross-lingual performance relative to LLMs. However, in few-shot cross-lingual settings, public LLMs demonstrate an enhanced adaptive potential. In addition, we observe that proprietary GPT-3.5 and GPT-4 lead in zero-shot cross-lingual capability, but are outpaced by public models in few-shot scenarios.