From Representations to Behaviors: Exploring the Person-Situation-Behavior Triad in LLMs

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Ruikang Zhang, Shuo Wang, Qi Su
• Venue: arXiv: cs.CL
• Published: 2026-07-29
• Collected: 2026-07-30
• URL: http://arxiv.org/abs/2607.26853v1
• PDF: https://arxiv.org/pdf/2607.26853v1
• DOI: UNKNOWN

Plain-Language Brief

The paper presents a framework for analyzing personality-related representations in Language Models (LLMs). The authors adapt Funder's personality triad framework to LLM analysis: Person as internal personality representations, Situation as contexts that trigger trait-relevant responses, and Behavior as response patterns on social tasks. They introduce a method for identifying sparse internal features associated with opposing poles of personality traits, validate their relevance, demonstrate consistent expression across contexts, and reveal behavioral changes in social intelligence tasks.

What the Document Says

• The paper discusses a framework for analyzing personality-related representations in LLMs.
• It adapts Funder's personality triad (Person, Situation, Behavior) for LLM analysis.
• They introduce a method for identifying internal features associated with opposing poles of personality traits using SAE decomposition and contrastive behavior pairs.
• The authors validate the trait relevance through effects on behavior to situation, token-level activation patterns, and robustness to paraphrasing.
• Feature-level interventions induce bidirectional trait-related shifts across a diverse set of situations while preserving response validity.
• The same interventions reveal behavioral changes in social intelligence tasks with benefit-tradeoff patterns consistent with human personality research.

Key Concepts Mentioned

• Language Models (LLMs)
• Personality Triad
• Personality-related internal representations
• Contrastive behavior pairs
• SAE decomposition
• Situation
• Behavior
• Feature-level interventions
• Social intelligence tasks
• Benefit-tradeoff patterns

Claims Made by the Document

• LLMs contain controllable trait-like representations linking internal states, situational expression, and behavioral outcomes.

Evidence Used

• Identification of sparse internal features associated with opposing poles of personality traits.
• Validation of trait relevance through effects on behavior to situation, token-level activation patterns, and robustness to paraphrasing.
• Demonstration of consistent expression across contexts through feature-level interventions.
• Revealing behavioral changes in social intelligence tasks with benefit-tradeoff patterns.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

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.