通过分步推理提升大模型偏见检测准确性
BiasGuard: A Reasoning-enhanced Bias Detection Tool For Large Language Models
- 分两阶段构建，先基于公平规范推理，再用强化学习优化判断能力
- 在五个数据集上准确率更高，过公平误判减少37%
- 适合需要高可信度偏见检测的AI伦理研究者与开发者
识别大语言模型生成内容中的偏见是确保其公平性的关键前提。现有方法如公平性分类器和基于大模型的评判工具，受限于难以理解潜在意图以及缺乏公平性判断标准。本文提出BiasGuard，一种新型偏见检测工具，通过显式分析输入并依据公平性规范进行推理，实现精准判断。该工具采用两阶段架构：第一阶段将模型初始化为基于公平性规范进行显式推理，第二阶段利用强化学习增强其推理与判断能力。实验在五个数据集上进行，结果表明BiasGuard优于现有工具，在准确率提升的同时显著减少过公平误判。研究还验证了增强推理对决策的重要性，并证明了两阶段优化流程的有效性。
原文摘要 · Abstract (English)
Identifying bias in LLM-generated content is a crucial prerequisite for ensuring fairness in LLMs. Existing methods, such as fairness classifiers and LLM-based judges, face limitations related to difficulties in understanding underlying intentions and the lack of criteria for fairness judgment. In this paper, we introduce BiasGuard, a novel bias detection tool that explicitly analyzes inputs and reasons through fairness specifications to provide accurate judgments. BiasGuard is implemented through a two-stage approach: the first stage initializes the model to explicitly reason based on fairness specifications, while the second stage leverages reinforcement learning to enhance its reasoning and judgment capabilities. Our experiments, conducted across five datasets, demonstrate that BiasGuard outperforms existing tools, improving accuracy and reducing over-fairness misjudgments. We also highlight the importance of reasoning-enhanced decision-making and provide evidence for the effectiveness of our two-stage optimization pipeline.
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。