Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Persona-Pruner: Sculpting Lightweight Models for Role-Playing

About

Language Models (LMs) have shown remarkable potential as role-playing chatbots, delivering consistent, stylized interactions when given a specification of a character or user persona. However, applying these capabilities to real-world applications (e.g., ecosystems with numerous NPCs interacting simultaneously) exposes a critical inefficiency due to the excessive computational cost. In this paper, we question the necessity of dedicating a full, generalist model to a single persona, hypothesizing that a specific character identity relies on only a fraction of the model's total capacity. We observe that naively pruning LMs often severely degrades the role-playing performance for a specific persona; it does not distinguish between redundant knowledge and essential character traits. We propose Persona-Pruner, a framework that sculpts a lightweight role-playing model by isolating persona-specific sub-networks from a single description. Our experiments consistently show that Persona-Pruner preserves role-playing performance substantially more effectively than existing state-of-the-art LLM pruning techniques, reducing the performance drop from the dense model by up to 93.8% over the strongest baseline on RoleBench in LLM-as-a-judge score, while still maintaining general LLM capabilities. Code is available at https://github.com/jsu-kim/Persona-Pruner.

Jinsu Kim, Jihoon Tack, Noah Lee, Jongheon Jeong• 2026

Related benchmarks

TaskDatasetResultRank
Question AnsweringPIQA
Accuracy75
589
Role-playingAlpaca-P
LLM-as-a-Judge Score83.9
91
Role-playing evaluationRoleBench
LLM-as-a-Judge Score85.67
44
General Knowledge QAOBQA
Normalized Accuracy40
44
General Knowledge QAPIQA
Accuracy73
44
Role-playingRoleBench (test)
LLM-as-a-Judge Score84.33
42
General CapabilityOBQA (test)
Normalized Accuracy40
42
Question AnsweringOBQA
Accuracy42
16
Language UnderstandingMMLU
Accuracy59.01
7
Role-playingAlpaca-P 50P x 100Q scale (test)
Alpaca-P Score81.97
5
Showing 10 of 13 rows

Other info

Follow for update