Are Large Language Models Capable of Generating Human-Level Narratives?
This addresses the problem of evaluating and enhancing narrative generation in LLMs for applications like creative writing and AI-assisted content creation, though it is incremental in improving existing methods.
This paper investigates the capability of large language models (LLMs) in storytelling, revealing that LLM-generated stories are homogeneously positive and lack tension compared to human-written ones, but explicit integration of discourse features improves neural storytelling by over 40% in diversity, suspense, and arousal.
This paper investigates the capability of LLMs in storytelling, focusing on narrative development and plot progression. We introduce a novel computational framework to analyze narratives through three discourse-level aspects: i) story arcs, ii) turning points, and iii) affective dimensions, including arousal and valence. By leveraging expert and automatic annotations, we uncover significant discrepancies between the LLM- and human- written stories. While human-written stories are suspenseful, arousing, and diverse in narrative structures, LLM stories are homogeneously positive and lack tension. Next, we measure narrative reasoning skills as a precursor to generative capacities, concluding that most LLMs fall short of human abilities in discourse understanding. Finally, we show that explicit integration of aforementioned discourse features can enhance storytelling, as is demonstrated by over 40% improvement in neural storytelling in terms of diversity, suspense, and arousal.