llm-editing/HalluEditBench
收藏Hugging Face2025-06-09 更新2025-04-26 收录
下载链接：
https://hf-mirror.com/datasets/llm-editing/HalluEditBench
下载链接
链接失效反馈官方服务：
资源简介：
HalluEditBench是一个全面评估知识编辑方法在纠正大型语言模型（LLMs）中虚假信息性能的数据集。它包含了9个领域、26个主题和超过6000个虚假信息实例，用于评估知识编辑方法在五个维度上的性能，包括有效性、泛化能力、迁移性、局部性和鲁棒性。
HalluEditBench is a dataset designed to holistically benchmark the performance of knowledge editing methods in correcting hallucinations in Large Language Models (LLMs). It contains over 6,000 hallucination instances across 9 domains and 26 topics, used to assess the performance of knowledge editing methods on five dimensions: Efficacy, Generalization, Portability, Locality, and Robustness.
提供机构：
llm-editing
