Does Mapo Tofu Contain Coffee? Probing LLMs for Food-related Cultural Knowledge

TL;DR

This paper investigates how large language models understand and exhibit biases in food-related cultural knowledge, introducing a multilingual dataset and analyzing factors influencing their cultural comprehension.

Contribution

It introduces FmLAMA, a multilingual dataset on food culture, and systematically evaluates LLMs' cultural knowledge across languages and architectures.

Findings

LLMs show bias towards US food knowledge

Cultural context improves LLM performance

Model effectiveness depends on language, architecture, and cultural nuances

Abstract

Recent studies have highlighted the presence of cultural biases in Large Language Models (LLMs), yet often lack a robust methodology to dissect these phenomena comprehensively. Our work aims to bridge this gap by delving into the Food domain, a universally relevant yet culturally diverse aspect of human life. We introduce FmLAMA, a multilingual dataset centered on food-related cultural facts and variations in food practices. We analyze LLMs across various architectures and configurations, evaluating their performance in both monolingual and multilingual settings. By leveraging templates in six different languages, we investigate how LLMs interact with language-specific and cultural knowledge. Our findings reveal that (1) LLMs demonstrate a pronounced bias towards food knowledge prevalent in the United States; (2) Incorporating relevant cultural context significantly improves LLMs' ability to access cultural knowledge; (3) The efficacy of LLMs in capturing cultural nuances is highly dependent on the interplay between the probing language, the specific model architecture, and the cultural context in question. This research underscores the complexity of integrating cultural understanding into LLMs and emphasizes the importance of culturally diverse datasets to mitigate biases and enhance model performance across different cultural domains.