Enregistré dans:
Détails bibliographiques
Auteurs principaux: Su, Yi, Zhang, Jiayi, Yang, Shu, Wang, Xinhai, Hu, Lijie, Wang, Di
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2505.17712
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915300974264320
author Su, Yi
Zhang, Jiayi
Yang, Shu
Wang, Xinhai
Hu, Lijie
Wang, Di
author_facet Su, Yi
Zhang, Jiayi
Yang, Shu
Wang, Xinhai
Hu, Lijie
Wang, Di
contents Rapid integration of large language models (LLMs) into societal applications has intensified concerns about their alignment with universal ethical principles, as their internal value representations remain opaque despite behavioral alignment advancements. Current approaches struggle to systematically interpret how values are encoded in neural architectures, limited by datasets that prioritize superficial judgments over mechanistic analysis. We introduce ValueLocate, a mechanistic interpretability framework grounded in the Schwartz Values Survey, to address this gap. Our method first constructs ValueInsight, a dataset that operationalizes four dimensions of universal value through behavioral contexts in the real world. Leveraging this dataset, we develop a neuron identification method that calculates activation differences between opposing value aspects, enabling precise localization of value-critical neurons without relying on computationally intensive attribution methods. Our proposed validation method demonstrates that targeted manipulation of these neurons effectively alters model value orientations, establishing causal relationships between neurons and value representations. This work advances the foundation for value alignment by bridging psychological value frameworks with neuron analysis in LLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2505_17712
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Understanding How Value Neurons Shape the Generation of Specified Values in LLMs
Su, Yi
Zhang, Jiayi
Yang, Shu
Wang, Xinhai
Hu, Lijie
Wang, Di
Computation and Language
Rapid integration of large language models (LLMs) into societal applications has intensified concerns about their alignment with universal ethical principles, as their internal value representations remain opaque despite behavioral alignment advancements. Current approaches struggle to systematically interpret how values are encoded in neural architectures, limited by datasets that prioritize superficial judgments over mechanistic analysis. We introduce ValueLocate, a mechanistic interpretability framework grounded in the Schwartz Values Survey, to address this gap. Our method first constructs ValueInsight, a dataset that operationalizes four dimensions of universal value through behavioral contexts in the real world. Leveraging this dataset, we develop a neuron identification method that calculates activation differences between opposing value aspects, enabling precise localization of value-critical neurons without relying on computationally intensive attribution methods. Our proposed validation method demonstrates that targeted manipulation of these neurons effectively alters model value orientations, establishing causal relationships between neurons and value representations. This work advances the foundation for value alignment by bridging psychological value frameworks with neuron analysis in LLMs.
title Understanding How Value Neurons Shape the Generation of Specified Values in LLMs
topic Computation and Language
url https://arxiv.org/abs/2505.17712