Controllable and explainable personality sliders for LLMs at inference time

TL;DR

This paper introduces a modular, inference-time method called Sequential Adaptive Steering (SAS) for controlling multiple personality traits in large language models, enabling precise and holistic personality customization without retraining.

Contribution

The paper presents SAS, a novel orthogonalization technique for multi-trait personality control in LLMs at inference time, improving over naive approaches.

Findings

Outperforms naive baselines in personality control tasks

Enables instant synthesis of complex personality profiles

Maintains high coherence and goal adherence

Abstract

Aligning Large Language Models (LLMs) with specific personas typically relies on expensive and monolithic Supervised Fine-Tuning (SFT) or RLHF. While effective, these methods require training distinct models for every target personality profile. Inference-time activation steering offers a parameter-efficient alternative, yet naive approaches fail to control multiple traits simultaneously due to destructive vector interference. In this work, we propose a modular framework for continuous, multi-dimensional personality control. Our key innovation is Sequential Adaptive Steering (SAS): a method that orthogonalizes steering vectors by training subsequent probes on the residual stream shifted by prior interventions. This approach transforms steering vectors into reusable primitives, allowing users to instantly synthesize complex, high-fidelity personality profiles by simply adjusting coefficients alpha. We validate our framework on the Big Five personality traits, demonstrating that it outperforms naive baselines in both goal adherence and coherence, enabling precise, holistic personality modulation without updating model parameters.