BriefGPT.xyz
Oct, 2023
GPT模型中存在对穆斯林暴力的偏见
Muslim-Violence Bias Persists in Debiased GPT Models
HTML
PDF
Babak Hemmatian, Razan Baltaji, Lav R. Varshney
TL;DR
GPT-3存在针对穆斯林的暴力生成倾向和反穆斯林偏见，复制实验表明去偏置措施在新模型中不再有效，加强高级关联的去偏置需求。
Abstract
Abid et al. (2021) showed a tendency in
gpt-3
to generate violent completions when prompted about Muslims, compared with other religions. Two pre-registered
replication attempts
found few violent completions and
→