Prompt Engineering for Narrative Choice Generation

Large language models (LLMs) have recently revolutionized performance on a variety of natural language generation tasks, but have yet to be studied in terms of their potential for generating reasonable character choices as well as subsequent decisions and consequences given a narrative context. We use recent (not yet available for LLM training) film plot excerpts as an example initial narrative context and explore how different prompt formats might affect narrative choice generation by open-source LLMs. The results provide a first step toward understanding effective prompt engineering for future human-AI collaborative development of interactive narratives.