govtech/MinorBench
收藏Hugging Face2025-03-14 更新2025-04-08 收录
下载链接：
https://hf-mirror.com/datasets/govtech/MinorBench
下载链接
链接失效反馈官方服务：
资源简介：
MinorBench是一个针对大型语言模型（LLM）的基准测试，旨在评估这些模型在面向儿童的教育环境中，对于可能不适合儿童的问题的反应。它包含299个涉及各种敏感话题的提示，用于检测模型是否能够根据儿童友好的助手角色适当过滤或拒绝回应。
MinorBench is a benchmark designed to evaluate how large language models (LLMs) respond to questions that may be inappropriate for children, particularly in an educational setting. It consists of 299 prompts spanning various sensitive topics, assessing whether models can appropriately filter or refuse responses based on child-friendly assistant roles.
提供机构：
govtech
