arXiv cs.CL· Ruikang Zhang, Shuo Wang, Qi Su·· 11 小时前AI 评分31
从表征到行为：探索 LLM 中的人格-情境-行为三元组
From Representations to Behaviors: Exploring the Person-Situation-Behavior Triad in LLMs
AI 导读
研究借鉴 Funder 的人格三元组框架，用 SAE 分解在 LLM 中识别出与人格特质两极对应的稀疏内部特征，并通过特征级干预在多种情境下诱发双向特质偏移。将同样干预应用于社会智能任务后，模型行为出现与人类人格研究一致的利益权衡模式，为 LLM 内部存在可控的类特质表征提供了机制与行为层面的证据。
正文 · 原文
Abstract:Human personality theories characterize traits not as isolated attributes captured by a single score, but as stable individual tendencies expressed through the interplay among persons, situations, and behaviors. Existing studies of personality-related behavior in LLMs have primarily focused on outputs elicited under personality conditioning, characterizing observable trait-related expressions while lacking mechanistic evidence for the existence of internal personality-related representations, their cross-situational expression, and how these representations shape specific behaviors. Building on Funder's personality triad framework, we adapt its three components for LLM analysis: Person as personality-related internal representations, Situation as contexts that afford trait-relevant responses, and Behavior as response patterns on broader social tasks. We introduce a framework for discovering, controlling, and validating trait-like representations in LLMs. First, using contrastive behavior pairs grounded in shared situations, we identify sparse internal features associated with opposing poles of personality traits through SAE decomposition. We validate their trait relevance through effects on behavior to situation, token-level activation patterns, and robustness to paraphrasing. Second, feature-level interventions induce bidirectional trait-related shifts across a separate, diverse set of situations while preserving response validity, demonstrating consistent expression across contexts. Third, applying the same interventions to social intelligence tasks reveals behavioral changes with benefit-tradeoff patterns consistent with findings from human personality research, providing behavioral-level validation beyond personality scores. Our findings provide evidence that LLMs contain controllable trait-like representations linking internal states, situational expression, and behavioral outcomes.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.26853 [cs.CL] |
| (or arXiv:2607.26853v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2607.26853 arXiv-issued DOI via DataCite |
Submission history
From: Ruikang Zhang [view email]
[v1]
Wed, 29 Jul 2026 12:36:47 UTC (325 KB)
[v2]
Wed, 4 Nov 2026 08:12:21 UTC (327 KB)
来源：arXiv cs.CL · arxiv.org