Chinese Large Language Models Evaluating and Enhancing Safety

Chinese Large Language Models are rapidly gaining popularity, with models such as ChatGPT (OpenAI, 2022) and GPT-4 (OpenAI, 2023) being widely used. However, these models also pose potential safety concerns such as generating insulting and discriminatory content, reflecting incorrect social values, and being used for malicious purposes such as fraud and dissemination of misleading information. In this paper, we evaluate and enhance the safety of Chinese Large Language Models, emphasizing the importance of responsible use and development of these technologies.