通过精准干预注意力头，定位并修复Transformer模型中的偏见。
Toward Localizing and Repairing Bias in Transformer Attention Heads
- 用公平性探测评估注意力头敏感度，定位偏见源头。
- 在4个模型上降低WinoBias差距，且语言质量更优。
- 适合需要可解释性修复的AI系统开发者。
Transformer语言模型被广泛用于软件组件，但其输出中的偏见仍难以定位与修复。现有公平性测试和修复方法多在输入-输出或重训练层面操作，而近期研究发现偏见行为可能集中于少数注意力头。本文探讨是否可通过针对性的推理时干预实现注意力头的定位与修复。提出ROBIN方法，一种白盒头级公平性调试工具，通过公平性探测的敏感度对注意力头排序，并从选定头的输出中移除一个小的偏见子空间。在四个模型的初步研究中，ROBIN在所有模型上均降低了测量的WinoBias差距，且比整体头置零更好地保持了语言建模质量。初步结果表明，头级偏见修复不仅需关注选择哪些头，还需考虑如何修改选定头。
原文摘要 · Abstract (English)
Transformer language models are increasingly used as software components, yet biased outputs remain difficult to localize and repair inside the model. Existing fairness testing and repair methods largely operate at the input-output or retraining level, while recent work suggests that bias-related behavior can concentrate in a small set of attention heads. This paper studies whether attention heads can be localized and repaired through a targeted inference-time intervention. We introduce ROBIN, a white-box head-level fairness debugging method that ranks attention heads using sensitivity to fairness probes and removes a small bias subspace from selected head outputs. In a four-model pilot study, ROBIN reduces the measured WinoBias gap across all models while preserving language-modeling quality better than whole-head zeroing. These preliminary results suggest that head-level bias repair should consider not only which heads are selected, but also how selected heads are modified.
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。