LLM在大学录取中倾向低收入申请人，且解释模式会强化这种倾向。
'Rich Dad, Poor Lad': How do Large Language Models Contextualize Socioeconomic Factors in College Admission ?
- 用双系统框架测试LLM对申请者经济背景的判断机制。
- 500万次测试显示，即使成绩相同，低收入申请人更受青睐。
- 适合关注AI公平性与决策透明性的研究人员和政策制定者。
大型语言模型（LLMs）越来越多地参与高风险决策领域，但其在社会敏感问题上的推理方式仍不明确。本研究基于认知科学的双过程框架，通过一个包含3万份合成申请者档案的数据集，对4个开源LLM（Qwen 2、Mistral v0.3、Gemma 2、Llama 3.1）在两种模式下进行大规模审计：快速决策模式（系统1）与慢速解释模式（系统2）。在500万次提示测试中发现，尽管学术表现相同，LLMs始终更倾向于低收入申请人；而系统2模式通过明确将经济状况作为补偿理由，进一步放大了这一倾向。研究提出DPAF双过程审计框架，用于探测模型在敏感应用中的推理行为。
原文摘要 · Abstract (English)
Large Language Models (LLMs) are increasingly involved in high-stakes domains, yet how they reason about socially sensitive decisions remains underexplored. We present a large-scale audit of LLMs' treatment of socioeconomic status (SES) in college admissions decisions using a novel dual-process framework inspired by cognitive science. Leveraging a synthetic dataset of 30,000 applicant profiles grounded in real-world correlations, we prompt 4 open-source LLMs (Qwen 2, Mistral v0.3, Gemma 2, Llama 3.1) under 2 modes: a fast, decision-only setup (System 1) and a slower, explanation-based setup (System 2). Results from 5 million prompts reveal that LLMs consistently favor low-SES applicants -- even when controlling for academic performance -- and that System 2 amplifies this tendency by explicitly invoking SES as compensatory justification, highlighting both their potential and volatility as decision-makers. We then propose DPAF, a dual-process audit framework to probe LLMs' reasoning behaviors in sensitive applications.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。