Attribution Bias in LLMs
收藏Mendeley Data2026-04-09 收录
下载链接：
https://data.mendeley.com/datasets/g65cxm6s99/1
下载链接
链接失效反馈官方服务：
资源简介：
A controlled experiment involving two event types, two attribution perspectives, four social attributes, and prompt-based interventions was conducted across five LLMs, including ChatGPT and four leading Chinese models.
