arXiv 2508.01710

CultureGuard: Towards Culturally-Aware Dataset and Guard Model for Multilingual Safety Applications

By Raviraj Joshi, Rakesh Paul, et al.

Published 2025-08-03

Mindmap

Browse the paper's core ideas, clusters, and relationships in a structured outline.

The increasing use of Large Language Models (LLMs) in agentic applications highlights the need for robust safety guard models. While content safety in English is well-studied, non-English languages lack similar advancements due to the high cost of collecting culturally aligned labeled datasets. We present CultureGuard, a novel solution for curating culturally aligned, high-quality safety datasets across multiple lan…

View the original paper on arXiv