自行去偏：教大语言模型认知偏差缓解干预措施
Debias It Yourself: Teaching LLMs Cognitive Bias Mitigation Interventions
- George Mason University（乔治梅森大学）
- University of Washington（华盛顿大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
提出基于认知的DIY框架，将五种人类去偏干预转化为大语言模型的去偏程序，通过展示、训练、修订三种范式实施，在多个基准上显著降低偏见并保持推理性能。
AI中文摘要：
偏见在社会心理学和认知科学中已被长期研究，数十年的研究产生了一系列经过验证的干预措施，这些措施能减少人类的刻板思维和偏见反应。我们提出“自行去偏”（DIY），一个基于认知的框架，将五种此类干预措施转化为大语言模型的去偏程序，并通过三种既定范式实施：展示（上下文示例）、训练（指令微调）和修订（引导式自我修订）。在三个模型、五个偏见基准、十一个去偏基线和三个推理基准上，训练+修订和仅修订取得了前两名的平均排名，在偏见-推理权衡中领先（在90%推理准确率下平均偏见低至2%），并在未见维度上将偏见最多降低14.8%。我们的代码和数据公开可用。
英文摘要：
Bias has long been studied in social psychology and cognitive science, where decades of research have produced a body of validated interventions that reduce stereotypical thinking and prejudiced responses in humans. We propose Debias It Yourself (DIY), a cognitively grounded framework that translates five such interventions into debiasing procedures for large language models and delivers them through three established paradigms: Show (in-context examples), Train (instruction tuning), and Revise (guided self-revision). Across three models, five bias benchmarks, eleven debiasing baselines, and three reasoning benchmarks, Train+Revise and Revise alone attain the top two average ranks, lead the bias-reasoning tradeoff (mean bias as low as 2% at 90% reasoning accuracy), and reduce bias on unseen dimensions by up to 14.8%. Our code and data are publicly available.
