ChineseSafe: A Chinese Benchmark for Evaluating Safety in Large Language Models

TL;DR

ChineseSafe is a comprehensive benchmark designed to evaluate the safety of large language models in Chinese, focusing on illegal and unsafe content detection, including politically sensitive and pornographic material, to improve content moderation and legal compliance.

Contribution

This work introduces ChineseSafe, a large Chinese safety benchmark with over 200,000 examples, addressing the lack of Chinese-specific safety evaluation for LLMs and including new illegal content categories.

Findings

Many LLMs are vulnerable to safety issues in Chinese contexts.

Current models pose legal risks due to safety vulnerabilities.

Benchmark results guide safer LLM development.

Abstract

With the rapid development of Large language models (LLMs), understanding the capabilities of LLMs in identifying unsafe content has become increasingly important. While previous works have introduced several benchmarks to evaluate the safety risk of LLMs, the community still has a limited understanding of current LLMs' capability to recognize illegal and unsafe content in Chinese contexts. In this work, we present a Chinese safety benchmark (ChineseSafe) to facilitate research on the content safety of large language models. To align with the regulations for Chinese Internet content moderation, our ChineseSafe contains 205,034 examples across 4 classes and 10 sub-classes of safety issues. For Chinese contexts, we add several special types of illegal content: political sensitivity, pornography, and variant/homophonic words. Moreover, we employ two methods to evaluate the legal risks of popular LLMs, including open-sourced models and APIs. The results reveal that many LLMs exhibit vulnerability to certain types of safety issues, leading to legal risks in China. Our work provides a guideline for developers and researchers to facilitate the safety of LLMs. Our results are also available at https://huggingface.co/spaces/SUSTech/ChineseSafe-Benchmark. Additionally, we release a test set comprising 200,000 examples, which is publicly accessible at https://huggingface.co/datasets/SUSTech/ChineseSafe.