“Now we know that we are living our lives in a book.”

“If what you say is true, I’m going to run away from the book and go my own way.”

In BookWorld, we create multi-agent societies for characters from fictional books, enabling them to engage in dynamic interactions that transcend their original narratives, thereby crafting innovative storytelling.

In this paper, we present CoSER, a collection of a high-quality dataset, open models, and an evaluation protocol towards effective RPLAs of established characters. The CoSER dataset covers 17,966 characters from 771 renowned books, providing authentic dialogues with real-world intricacies & diverse data types such as conversation setups, character experiences and internal thoughts. Drawing from acting methodology, we introduce given-circumstance acting for training and evaluating role-playing LLMs, where LLMs sequentially portray multiple characters in book scenes. We release CoSER 8B and CoSER 70B, i.e., state-of-the-art open role-playing LLMs built on LLaMA-3.1.

We present a comprehensive survey on role-playing language agents (RPLAs), i.e., specialized AI systems simulating assigned personas. Specifically, we distinguish personas in RPLAs into three progressive tiers - demographic persona, character persona, and individualized persona. Our survey discusses their data sourcing, agent construction, evaluation, applications, risks, limitations and future prospects.

Do role-playing language agents accurately capture character personalities? Towards this question, we propose InCharacter that evaluates their personality fidelity via psychological interviews using 14 scales.

Can Large Language Models substitute humans in making important decisions? Recent research has unveiled the potential of LLMs to role-play assigned personas, mimicking their knowledge and linguistic habits. However, imitative decision-making requires a more nuanced understanding of personas. In this paper, we benchmark the ability of LLMs in persona-driven decision-making. Specifically, we investigate whether LLMs can predict characters’ decisions provided with the preceding stories in high-quality novels. Leveraging character analyses written by literary experts, we construct a dataset LIFECHOICE comprising 1,401 character decision points from 395 books. Then, we conduct comprehensive experiments on LIFECHOICE, with various LLMs and methods for LLM role-playing. The results demonstrate that state-of-the-art LLMs exhibit promising capabilities in this task, yet there is substantial room for improvement. Hence, we further propose the CHARMAP method, which achieves a 6.01% increase in accuracy via persona-based memory retrieval. We will make our datasets and code publicly available.