Researchers have developed WikiOriginQA, a new question-answering dataset automatically constructed using the creation dates of Wikipedia language editions to represent knowledge origins. This method aims to overcome the scalability challenges of manual annotation for cultural knowledge datasets. Experiments using WikiOriginQA indicate that large language models exhibit bias towards Anglophone cultures and highlight the significance of knowledge propagation into the Anglosphere. AI
IMPACT Highlights potential biases in LLMs, suggesting a need for more culturally diverse training data and evaluation methods.
RANK_REASON The cluster contains an academic paper detailing a new dataset and its findings. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
