Adapting Safe-for-Work Classifier for Malaysian Language Text: Enhancing Alignment in LLM-Ops Framework

TL;DR

This paper introduces a novel safe-for-work classifier tailored for Malaysian language content, addressing a critical gap in content safety for multilingual LLM deployment, and provides a publicly available model to enhance alignment in LLM-Ops.

Contribution

The paper presents the first Malaysian language safe-for-work classifier, including a curated dataset and a trained model, to improve content safety in multilingual LLM applications.

Findings

Developed a Malaysian safe-for-work classifier with high accuracy.

Curated and annotated a diverse Malaysian language dataset.

Publicly released the classifier for community use.

Abstract

As large language models (LLMs) become increasingly integrated into operational workflows (LLM-Ops), there is a pressing need for effective guardrails to ensure safe and aligned interactions, including the ability to detect potentially unsafe or inappropriate content across languages. However, existing safe-for-work classifiers are primarily focused on English text. To address this gap for the Malaysian language, we present a novel safe-for-work text classifier tailored specifically for Malaysian language content. By curating and annotating a first-of-its-kind dataset of Malaysian text spanning multiple content categories, we trained a classification model capable of identifying potentially unsafe material using state-of-the-art natural language processing techniques. This work represents an important step in enabling safer interactions and content filtering to mitigate potential risks and ensure responsible deployment of LLMs. To maximize accessibility and promote further research towards enhancing alignment in LLM-Ops for the Malaysian context, the model is publicly released at https://huggingface.co/malaysia-ai/malaysian-sfw-classifier.