Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Drama Llama: An LLM-Powered Storylets Framework for Authorable Responsiveness in Interactive Narrative

Yuqian Sun, Phoebe J. Wang, John Joon Young Chung, Melissa Roemmele, Taewook Kim, Max Kreminski

TL;DR

Drama Llama tackles the tension between authorial control and emergent narrative in interactive storytelling by uniting storylet-driven content with LLM-based generation. The framework lets authors write world settings, character prompts, and natural-language triggers that govern when narrative events occur, while LLM agents drive character behavior and dialogue within those constraints. In a preliminary study with six authors, the system produced coherent narratives and believable interactions, though originality and character depth were uneven and depended on author input. The work demonstrates that a hybrid approach can provide structured authorial influence and flexible generation, offering a practical path for scalable, responsive interactive narratives.

Abstract

In this paper, we present Drama Llama, an LLM-powered storylets framework that supports the authoring of responsive, open-ended interactive stories. DL combines the structural benefits of storylet-based systems with the generative capabilities of large language models, enabling authors to create responsive interactive narratives while maintaining narrative control. Rather than crafting complex logical preconditions in a general-purpose or domain-specific programming language, authors define triggers in natural language that fire at appropriate moments in the story. Through a preliminary authoring study with six content authors, we present initial evidence that DL can generate coherent and meaningful narratives with believable character interactions. This work suggests directions for hybrid approaches that enhance authorial control while supporting emergent narrative generation through LLMs.

Drama Llama: An LLM-Powered Storylets Framework for Authorable Responsiveness in Interactive Narrative

TL;DR

Drama Llama tackles the tension between authorial control and emergent narrative in interactive storytelling by uniting storylet-driven content with LLM-based generation. The framework lets authors write world settings, character prompts, and natural-language triggers that govern when narrative events occur, while LLM agents drive character behavior and dialogue within those constraints. In a preliminary study with six authors, the system produced coherent narratives and believable interactions, though originality and character depth were uneven and depended on author input. The work demonstrates that a hybrid approach can provide structured authorial influence and flexible generation, offering a practical path for scalable, responsive interactive narratives.

Abstract

In this paper, we present Drama Llama, an LLM-powered storylets framework that supports the authoring of responsive, open-ended interactive stories. DL combines the structural benefits of storylet-based systems with the generative capabilities of large language models, enabling authors to create responsive interactive narratives while maintaining narrative control. Rather than crafting complex logical preconditions in a general-purpose or domain-specific programming language, authors define triggers in natural language that fire at appropriate moments in the story. Through a preliminary authoring study with six content authors, we present initial evidence that DL can generate coherent and meaningful narratives with believable character interactions. This work suggests directions for hybrid approaches that enhance authorial control while supporting emergent narrative generation through LLMs.
Paper Structure (17 sections, 5 figures, 2 tables)

This paper contains 17 sections, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. System Description
  4. Evaluation
  5. Task A: Open-Ended Authoring
  6. Task B: Fixed-Topic Authoring
  7. Results
  8. Torrance Tests for Creative Writing (TTCW)
  9. Author Feedback
  10. Comparison of Authored Stories
  11. Discussion
  12. Limitations and Future Work
  13. Conclusion
  14. Example Trigger Definition
  15. System Prompt for Simulation
  16. ...and 2 more sections

Figures (5)

  • Figure 1: Overall diagram of the Drama Llama system.
  • Figure 2: The setting configuration interface used by authors to define their story setting, characters, and triggers.
  • Figure 3: The simulation interface used by authors to test their story definition as they write and modify the setting and trigger content.
  • Figure 4: Textual and visual summaries of the stories produced for Task A by our six authoring study participants.
  • Figure 5: Authors' TTCW evaluations of representative playthroughs of their Task A and Task B stories. For both tasks, authors answered either "yes" or "no" to each question in the TTCW questionnaire; presented numbers indicate how many authors answered "yes" or "no" for each question and task.