A Systematic Analysis of the Impact of Persona Steering on LLM Capabilities

TL;DR

This paper systematically examines how inducing specific personas in LLMs affects their cognitive abilities, revealing stable, trait-dependent performance shifts and proposing a dynamic routing strategy to optimize outcomes.

Contribution

It introduces the NPTI framework for persona induction in LLMs, analyzes its effects on cognition, and proposes DPR for adaptive persona-based querying.

Findings

Persona induction causes stable shifts in cognitive performance.

Effects vary systematically by personality trait dimensions.

DPR outperforms static personas without extra training.

Abstract

Imbuing Large Language Models (LLMs) with specific personas is prevalent for tailoring interaction styles, yet the impact on underlying cognitive capabilities remains unexplored. We employ the Neuron-based Personality Trait Induction (NPTI) framework to induce Big Five personality traits in LLMs and evaluate performance across six cognitive benchmarks. Our findings reveal that persona induction produces stable, reproducible shifts in cognitive task performance beyond surface-level stylistic changes. These effects exhibit strong task dependence: certain personalities yield consistent gains on instruction-following, while others impair complex reasoning. Effect magnitude varies systematically by trait dimension, with Openness and Extraversion exerting the most robust influence. Furthermore, LLM effects show 73.68% directional consistency with human personality-cognition relationships. Capitalizing on these regularities, we propose Dynamic Persona Routing (DPR), a lightweight query-adaptive strategy that outperforms the best static persona without additional training.