作者位次 #4/9ACL 2026Findings· Findings of ACL
Debiasing LLM s by Masking Unfairness-Driving Attention Heads
Tingxu Han · Wei Song · Ziqi Ding · Ziming Li · Chunrong Fang · Yuekang Li · Dongfang Liu · Zhenyu Chen · Zhenting Wang
摘要
Large language models (LLMs) increasingly mediate decisions in domains where unfair treatment of demographic groups is unacceptable. Existing work probes when biased outputs appear, but gives little insight into the mechanisms that generate them, leaving existing mitigations largely fragile. In t...