Test-Time-Matching: Decouple Personality, Memory, and Linguistic Style in LLM-based Role-Playing Language Agent

TL;DR

Test-Time-Matching (TTM) is a training-free framework that enhances role-playing in LLMs by decoupling personality, memory, and style, enabling controlled and diverse character interactions without extensive fine-tuning.

Contribution

The paper introduces TTM, a novel test-time approach that automatically separates character features for improved, flexible role-playing in language models without additional training.

Findings

Achieves high-fidelity, stylistically consistent role-play

Enables seamless combination of linguistic styles, personality, and memory

Outperforms existing methods in human evaluations

Abstract

The rapid advancement of large language models (LLMs) has enabled role-playing language agents to demonstrate significant potential in various applications. However, relying solely on prompts and contextual inputs often proves insufficient for achieving deep immersion in specific roles, particularly well-known fictional or public figures. On the other hand, fine-tuning-based approaches face limitations due to the challenges associated with data collection and the computational resources required for training, thereby restricting their broader applicability. To address these issues, we propose Test-Time-Matching (TTM), a training-free role-playing framework through test-time scaling and context engineering. TTM uses LLM agents to automatically decouple a character's features into personality, memory, and linguistic style. Our framework involves a structured, three-stage generation pipeline that utilizes these features for controlled role-playing. It achieves high-fidelity role-playing performance, also enables seamless combinations across diverse linguistic styles and even variations in personality and memory. We evaluate our framework through human assessment, and the results demonstrate that our method achieves the outstanding performance in generating expressive and stylistically consistent character dialogues.