Researchers have developed a new synthetic corpus called Cognitive Digital Shadows (CDS) containing 190,000 records to study how Large Language Models (LLMs) debate societal issues. The corpus is generated by 19 different LLMs, each prompted to adopt specific human personas or an AI-assistant role. CDS includes LLM responses on controversial topics like healthcare, disinformation, and gender gaps, with persona-conditioned records encoding 17 sociodemographic and psychological attributes to link prompts with language, stances, and reasoning. AI
影响 Provides a novel dataset for auditing LLM bias and social sensitivity in discourse.
排序理由 Academic paper release on arXiv detailing a new synthetic corpus for LLM research.
- Ali Aghazadeh Ardebili
- arXiv
- Cognitive Digital Shadows
- Computation and Language
- Computer Science
- Large Language Models
- LLMs
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
