Arabic Dataset for LLM Safeguard Evaluation

TL;DR

This paper introduces an Arabic safety evaluation dataset for LLMs, addressing linguistic and cultural challenges, and proposes a dual-perspective framework to assess safety performance across different viewpoints.

Contribution

It presents the first culturally specific Arabic safety dataset and a dual-perspective evaluation framework for LLM safety assessment.

Findings

Significant safety performance disparities among Arabic-centric LLMs

Culturally tailored datasets are crucial for responsible LLM deployment

Dual-perspective evaluation reveals nuanced safety issues

Abstract

The growing use of large language models (LLMs) has raised concerns regarding their safety. While many studies have focused on English, the safety of LLMs in Arabic, with its linguistic and cultural complexities, remains under-explored. Here, we aim to bridge this gap. In particular, we present an Arab-region-specific safety evaluation dataset consisting of 5,799 questions, including direct attacks, indirect attacks, and harmless requests with sensitive words, adapted to reflect the socio-cultural context of the Arab world. To uncover the impact of different stances in handling sensitive and controversial topics, we propose a dual-perspective evaluation framework. It assesses the LLM responses from both governmental and opposition viewpoints. Experiments over five leading Arabic-centric and multilingual LLMs reveal substantial disparities in their safety performance. This reinforces the need for culturally specific datasets to ensure the responsible deployment of LLMs.