Personality Expression Across Contexts: Linguistic and Behavioral Variation in LLM Agents
This research addresses the problem of inconsistent personality expression in LLM-based dialogue agents, which is important for developers and users seeking reliable human-like interactions, though it is incremental in exploring context-sensitive adaptation.
The study investigated how identical personality prompts in LLMs lead to varied linguistic, behavioral, and emotional outcomes across different conversational contexts, revealing that contextual cues systematically influence personality expression and emotional tone.
Large Language Models (LLMs) can be conditioned with explicit personality prompts, yet their behavioral realization often varies depending on context. This study examines how identical personality prompts lead to distinct linguistic, behavioral, and emotional outcomes across four conversational settings: ice-breaking, negotiation, group decision, and empathy tasks. Results show that contextual cues systematically influence both personality expression and emotional tone, suggesting that the same traits are expressed differently depending on social and affective demands. This raises an important question for LLM-based dialogue agents: whether such variations reflect inconsistency or context-sensitive adaptation akin to human behavior. Viewed through the lens of Whole Trait Theory, these findings highlight that LLMs exhibit context-sensitive rather than fixed personality expression, adapting flexibly to social interaction goals and affective conditions.