arXiv 2604.279292026-04-30cs.CL

DPN-LE: 大语言模型双人格神经元定位与编辑

DPN-LE: Dual Personality Neuron Localization and Editing for Large Language Models

Southeast University（东南大学）；Shanghai Jiao Tong University（上海交通大学）；East China Normal University（华东师范大学）；Zhongguancun Academy（中关村学院）；Zhejiang University of Technology（浙江工业大学）；Kuaishou Technology（快手科技）；Northeastern University（东北大学）；Tsinghua University（清华大学）；Nanjing University of Aeronautics and Astronautics（南京航空航天大学）

Lifan Zheng, Xue Yang, Jiawei Chen, Chenyan Wu, Jingyuan Zhang, Fanheng Kong, Xinyi Zeng, Xiang Chen, Yu Tian

发表信息：: ACL 2026 Findings

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出DPN-LE方法，通过对比高特质与低特质样本的MLP激活，定位并编辑双人格神经元，实现精准的人格控制与能力保留。

摘要：

随着大型语言模型（LLMs）的广泛应用，理解其人格表示机制变得至关重要。作为人格编辑的一种新范式，现有方法多采用神经元编辑来定位和修改LLM神经元，需改变大量神经元，导致显著的性能下降。本文通过评估通用能力影响和表示层面模式，探讨并量化这种特异性。研究发现：1）现有方法可改变人格但降低整体性能；2）神经元多功能，连接人格特质与通用知识；3）对立人格特质表现出明显互斥的表示模式。受此启发，我们提出DPN-LE（双人格神经元定位与编辑），通过对比高特质与低特质样本的MLP激活，识别人格特定神经元。DPN-LE构建层间引导向量，并基于Cohen的$d$效应大小和激活幅度进行双准则过滤，以隔离互斥的神经元子集。对这些神经元的稀疏线性干预在推理任务中实现了精确的人格控制。使用每个特质1,000对对比样本，DPN-LE仅干预约0.5%的神经元，实现竞争性的人格控制和显著更好的能力保留。在LLaMA-3-8B-Instruct和Qwen2.5-7B-Instruct上的实验验证了该方法的有效性和通用性。

英文摘要：

With the widespread adoption of large language models (LLMs), understanding their personality representation mechanisms has become critical. As a novel paradigm in Personality Editing, most existing methods employ neuron-editing to locate and modify LLM neurons, requiring changes to numerous neurons and leading to significant performance degradation. This raises a fundamental question: Are all modified neurons directly related to personality representation? In this work, we investigate and quantify this specificity through assessments of general capability impact and representation-level patterns. We find that: 1) Current methods can change personalities but reduce overall performance. 2) Neurons are multifunctional, connecting personality traits and general knowledge. 3) Opposing personality traits demonstrate distinctly mutually exclusive representation patterns. Motivated by these findings, we propose DPN-LE (Dual Personality Neuron Localization and Editing), which identifies personality-specific neurons by contrasting MLP activations between high-trait and low-trait samples. DPN-LE constructs layer-wise steering vectors and applies dual-criterion filtering based on Cohen's $d$ effect size and activation magnitude to isolate mutually exclusive neuron subsets. Sparse linear intervention on these neurons enables precise personality control at inference time. Using only 1,000 contrastive sample pairs per trait, DPN-LE intervenes on $\sim$0.5\% of neurons while achieving competitive personality control and substantially better capability preservation across reasoning tasks. Experiments on LLaMA-3-8B-Instruct and Qwen2.5-7B-Instruct demonstrate the effectiveness and generalizability of our approach.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。