Abstract

Automated long-form story generation typically employs long-context largelanguage models (LLMs) for one-shot creation, which can produce cohesive butnot necessarily engaging content. We introduce Storytelling With ActionGuidance (SWAG), a novel approach to storytelling with LLMs. Our approachframes story writing as a search problem through a two-model feedback loop: oneLLM generates story content, and another auxiliary LLM is used to choose thenext best "action" to steer the story's future direction. Our results show thatSWAG can substantially outperform previous end-to-end story generationtechniques when evaluated by GPT-4 and through human evaluation. Our SWAGpipeline using only small open-source models surpasses GPT-3.5-Turbo.

 

Quick Read (beta)