StoryBox: Collaborative Multi-Agent Simulation for Hybrid Bottom-Up Long-Form Story Generation Using Large Language Models

TL;DR

StoryBox tackles long-form story generation by coupling emergent events from a collaborative multi-agent sandbox with a Storyteller that crafts coherent narratives from these events. The framework adopts a hybrid bottom-up approach where agent-driven interactions in a flexible tree-like environment surface rich events, which are summarized and fed into an information-retrieval driven generation loop to produce chapters that can exceed 10,000 words with sustained coherence. Key contributions include a Persona Scratch Information model, a scalable environment hierarchy, an event-centric generation pipeline, and extensive hybrid evaluations demonstrating state-of-the-art performance across coherence, character development, language use, and narrative depth. The work offers a scalable, immersive approach to dynamic storytelling with potential applications in interactive fiction, game design, and narrative research.

Abstract

Human writers often begin their stories with an overarching mental scene, where they envision the interactions between characters and their environment. Inspired by this creative process, we propose a novel approach to long-form story generation, termed hybrid bottom-up long-form story generation, using multi-agent simulations. In our method, agents interact within a dynamic sandbox environment, where their behaviors and interactions with one another and the environment generate emergent events. These events form the foundation for the story, enabling organic character development and plot progression. Unlike traditional top-down approaches that impose rigid structures, our hybrid bottom-up approach allows for the natural unfolding of events, fostering more spontaneous and engaging storytelling. The system is capable of generating stories exceeding 10,000 words while maintaining coherence and consistency, addressing some of the key challenges faced by current story generation models. We achieve state-of-the-art performance across several metrics. This approach offers a scalable and innovative solution for creating dynamic, immersive long-form stories that evolve organically from agent-driven interactions.

Figures (8)

• Figure 1: The timeline of the multi-agent sandbox simulation, where agent interactions with each other and their environment trigger emergent events that drive dynamic, hybrid bottom-up story generation.
• Figure 2: Overview of the system framework for long-form story generation, including the Persona Scratch Information for defining character settings, the sandbox where agent interactions generate events, and the Storyteller Agent that uses these events to craft a complete story.
• Figure 3: Overview of the Storyteller Agent workflow for generating long-form story using a hybrid bottom-up approach, from sandbox events to iterative story generation.
• Figure 4: Comparative performance of different methods across multi evaluation dimensions: Plot, Creativity, Character Development, Language Use, Conflict Quality, and Overall. Subfigure (a) presents rankings based on LLM-based evaluation, while subfigure (b) shows rankings from human evaluation. We also include the sandbox-specific metric Character Behavior Consistency, along with the Average Word Count for each method.
• Figure 5: Effect of simulation duration on story generation performance.