Recursive Reprompting and Story Generation
What topic would you be interested in seeing an article/chapter about?
- https://arxiv.org/abs/2210.06774
(a) prompting a general-purpose language model to construct a structured overarching plan, and (b) generating story passages by repeatedly injecting contextual information from both the plan and current story state into a language model prompt.
- https://arxiv.org/abs/2212.10077
DOC consists of two complementary components: a detailed outliner and a detailed controller. The detailed outliner creates a more detailed, hierarchically structured outline, shifting creative burden from the main drafting procedure to the planning stage. The detailed controller ensures the more detailed outline is still respected during generation by controlling story passages to align with outline details
Any additional info you have about it:
It might be interesting to write the entry about that technique in a more general way, not so focused on stories per se, but as an example of how an LLM's "narrative" can be controlled and refined.