VocalNet/VocalBench-zh
收藏Hugging Face2025-10-09 更新2025-11-01 收录
下载链接：
https://hf-mirror.com/datasets/VocalNet/VocalBench-zh
下载链接
链接失效反馈官方服务：
资源简介：
VocalBench-zh是一个用于评估多模态大型语言模型普通话语音交互能力的全面基准。该数据集包括中文知识、外语知识、通用知识等多个评价集，每个评价集包含1000个实例，涉及不同的数据资源。此外，还包括推理、创造力、单轮对话、多轮对话、安全性、情感同理心、代码切换和鲁棒性等评价集。数据集的许可为Apache-2.0。
VocalBench-zh is a comprehensive benchmark for evaluating the Mandarin Speech Interaction capabilities of multi-modal LLMs. The dataset includes Chinese Knowledge, Foreign Knowledge, General Knowledge, and other evaluation sets, each containing 1000 instances from various data resources. Additionally, it includes evaluation sets for Reasoning, Creativity, Single-Round, Multi-Round, Safety, Emotional Empathy, Code-Switching, and Robustness. The dataset is licensed under Apache-2.0.
提供机构：
VocalNet
