Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Fusian: Multi-LoRA Fusion for Fine-Grained Continuous MBTI Personality Control in Large Language Models

Zehao Chen, Rong Pan

Abstract

Large Language Models (LLMs) have demonstrated impressive capabilities in simulating diverse human behaviors and personalities. However, existing methods for personality control, which include prompt engineering and standard Supervised Fine-Tuning (SFT), typically treat personality traits as discrete categories (e.g., "Extroverted" vs. "Introverted"), lacking the ability to precisely control the intensity of a trait on a continuous spectrum. In this paper, we introduce Fusian, a novel framework for fine-grained, continuous personality control in LLMs. Fusian operates in two stages: (1) Trajectory Collection, where we capture the dynamic evolution of personality adoption during SFT by saving a sequence of LoRA adapters, effectively mapping the continuous manifold of a trait; and (2) RL-based Dynamic Fusion, where we train a policy network using Reinforcement Learning to dynamically compute mixing weights for these frozen adapters. By sampling from a Dirichlet distribution parameterized by the policy network, Fusian fuses multiple adapters to align the model's output with a specific numerical target intensity. Experiments on the Qwen3-14B model demonstrate that Fusian achieves high precision in personality control, significantly outperforming baseline methods in aligning with user-specified trait intensities.

Fusian: Multi-LoRA Fusion for Fine-Grained Continuous MBTI Personality Control in Large Language Models

Abstract

Large Language Models (LLMs) have demonstrated impressive capabilities in simulating diverse human behaviors and personalities. However, existing methods for personality control, which include prompt engineering and standard Supervised Fine-Tuning (SFT), typically treat personality traits as discrete categories (e.g., "Extroverted" vs. "Introverted"), lacking the ability to precisely control the intensity of a trait on a continuous spectrum. In this paper, we introduce Fusian, a novel framework for fine-grained, continuous personality control in LLMs. Fusian operates in two stages: (1) Trajectory Collection, where we capture the dynamic evolution of personality adoption during SFT by saving a sequence of LoRA adapters, effectively mapping the continuous manifold of a trait; and (2) RL-based Dynamic Fusion, where we train a policy network using Reinforcement Learning to dynamically compute mixing weights for these frozen adapters. By sampling from a Dirichlet distribution parameterized by the policy network, Fusian fuses multiple adapters to align the model's output with a specific numerical target intensity. Experiments on the Qwen3-14B model demonstrate that Fusian achieves high precision in personality control, significantly outperforming baseline methods in aligning with user-specified trait intensities.
Paper Structure (45 sections, 5 equations, 3 figures, 3 tables)

This paper contains 45 sections, 5 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. LLM-based Role-Playing Agents
  4. Personality Modeling and Control
  5. Methodology
  6. Stage 1: Trajectory Collection
  7. Supervised Fine-Tuning Setup
  8. High-Frequency Checkpointing and Evaluation
  9. Trajectory Library
  10. Stable Basis Selection
  11. Stage 2: RL-based Dynamic Fusion
  12. Policy Network Architecture
  13. Action and Fusion
  14. Reward Shaping and Optimization
  15. Experiments
  16. ...and 30 more sections

Figures (3)

  • Figure 1: Conceptual illustration of Fusian. Unlike discrete categorization, Fusian achieves continuous intensity control (e.g., 78% Perceiving) by dynamically fusing LoRA adapters via a policy network.
  • Figure 2: Overview of the framework. Stage 1 collects LoRA adapters during SFT, evaluating them via MBTI Test to get the personality manifold. Stage 2 trains a policy network via RL to dynamically fuse these adapters to match a target intensity. During Inference, the fused adapter generates responses aligned with the user's specification.
  • Figure 3: The personality manifold from the SFT trajectory. Trait intensity evolution reveals non-monotonicity and diminishing returns.