Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Orca: Enhancing Role-Playing Abilities of Large Language Models by Integrating Personality Traits

Yuxuan Huang

TL;DR

This paper proposes Orca, a framework for data processing and training LLMs of custom characters by integrating personality traits, and introduces OrcaBench, the first benchmark for evaluating the quality of content generated by LLMs on social platforms across multiple scales.

Abstract

Large language models has catalyzed the development of personalized dialogue systems, numerous role-playing conversational agents have emerged. While previous research predominantly focused on enhancing the model's capability to follow instructions by designing character profiles, neglecting the psychological factors that drive human conversations. In this paper, we propose Orca, a framework for data processing and training LLMs of custom characters by integrating personality traits. Orca comprises four stages: (1) Personality traits inferring, leverage LLMs to infer user's BigFive personality trait reports and scores. (2) Data Augment, simulate user's profile, background story, and psychological activities. (3) Dataset construction, personality-conditioned instruction prompting (PCIP) to stimulate LLMs. (4) Modeling and Training, personality-conditioned instruction tuning (PTIT and PSIT), using the generated data to enhance existing open-source LLMs. We introduce OrcaBench, the first benchmark for evaluating the quality of content generated by LLMs on social platforms across multiple scales. Our experiments demonstrate that our proposed model achieves superior performance on this benchmark, demonstrating its excellence and effectiveness in perceiving personality traits that significantly improve role-playing abilities. Our Code is available at https://github.com/Aipura/Orca.

Orca: Enhancing Role-Playing Abilities of Large Language Models by Integrating Personality Traits

TL;DR

This paper proposes Orca, a framework for data processing and training LLMs of custom characters by integrating personality traits, and introduces OrcaBench, the first benchmark for evaluating the quality of content generated by LLMs on social platforms across multiple scales.

Abstract

Large language models has catalyzed the development of personalized dialogue systems, numerous role-playing conversational agents have emerged. While previous research predominantly focused on enhancing the model's capability to follow instructions by designing character profiles, neglecting the psychological factors that drive human conversations. In this paper, we propose Orca, a framework for data processing and training LLMs of custom characters by integrating personality traits. Orca comprises four stages: (1) Personality traits inferring, leverage LLMs to infer user's BigFive personality trait reports and scores. (2) Data Augment, simulate user's profile, background story, and psychological activities. (3) Dataset construction, personality-conditioned instruction prompting (PCIP) to stimulate LLMs. (4) Modeling and Training, personality-conditioned instruction tuning (PTIT and PSIT), using the generated data to enhance existing open-source LLMs. We introduce OrcaBench, the first benchmark for evaluating the quality of content generated by LLMs on social platforms across multiple scales. Our experiments demonstrate that our proposed model achieves superior performance on this benchmark, demonstrating its excellence and effectiveness in perceiving personality traits that significantly improve role-playing abilities. Our Code is available at https://github.com/Aipura/Orca.

Paper Structure

This paper contains 19 sections, 1 equation, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Persona-based Dialogue
  4. Personality-aware Dialogue
  5. Character-based Dialogue
  6. Methods
  7. Personality traits inferring
  8. Data Augment
  9. Dataset Construction
  10. Model
  11. OrcaBench
  12. Experiment
  13. Implement Details
  14. Baselines
  15. Evaluation Metrics
  16. ...and 4 more sections

Figures (2)

  • Figure 1: The workflow for developing our personalized agent system, Orca, to provide personalized interaction on social media platforms. Orca comprises four stages: (1) Personality traits inferring; (2) Data Augment through designing numerous simulation prompts. (3) Dataset construction, serial instruction for the connection of labels and character and personality traits. (4) Modeling and Training, personality traits instruction tuning (PTIT) and personality scores tuning (PSIT), using the generated data to enhance existing open-source LLMs.
  • Figure 2: An example interaction between an personalized agent object of Orca and human on social platform. The bold blue text in the bubble indicates the correlation between the agent's psychological activities and personality traits.