stefan-it/nanochat-german-eval-data
收藏Hugging Face2025-10-21 更新2025-10-25 收录
下载链接:
https://hf-mirror.com/datasets/stefan-it/nanochat-german-eval-data
下载链接
链接失效反馈官方服务:
资源简介:
nanochat German: Evaluation Data是一个德语评估数据集,用于评估德语nanochat模型。数据集包括Commonsense Reasoning (COPA)、Language Understanding (HellaSwag)、Reading Comprehension (BoolQ)、Safety (Enterprise PII Classification)和World Knowledge (MMLU)五个部分,涵盖了常识推理、语言理解、阅读理解、安全和个人身份信息分类以及世界知识等领域。这些数据集是通过Gemini 2.5 Pro翻译工具从原始英文数据集翻译而来,用于评估模型的性能。
nanochat German: Evaluation Data is a German evaluation dataset used for assessing German nanochat models. The dataset includes five parts: Commonsense Reasoning (COPA), Language Understanding (HellaSwag), Reading Comprehension (BoolQ), Safety (Enterprise PII Classification), and World Knowledge (MMLU), covering areas such as common sense reasoning, language understanding, reading comprehension, safety, and personal identity information classification. These datasets have been translated from the original English datasets using the Gemini 2.5 Pro translation tool to evaluate the performance of the models.
提供机构:
stefan-it



