Summary of Evaluating Creative Short Story Generation in Humans and Large Language Models, by Mete Ismayilzada et al.
Evaluating Creative Short Story Generation in Humans and Large Language Models
by Mete Ismayilzada, Claire Stevenson, Lonneke van der Plas
First submitted to arxiv on: 4 Nov 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper investigates the creative capabilities of large language models (LLMs) in short story generation, comparing them with human writers. The authors use a five-sentence creative writing task and automate metrics to evaluate model- and human-generated stories across dimensions like novelty, surprise, diversity, and linguistic complexity. Results show that LLMs produce stylistically complex stories but struggle with novelty, surprise, and diversity compared to average human writers. Expert ratings align with automated metrics, while non-expert raters perceive LLM-generated stories as more creative than human-written ones. This study explores the differences in creativity ratings between humans and machines and discusses implications for both human and artificial creativity. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research looks at how well computer models can write short stories compared to people. The authors tested 60 computer models and 60 people by having them each write a five-sentence story. They used special tools to measure the stories’ creativity, including things like how unique and surprising they were. The results showed that while the computer models wrote very complex and interesting stories, they didn’t do as well as humans when it came to coming up with new ideas or surprising readers. People thought the computer-written stories were actually more creative than the ones written by people! This study helps us understand why machines and humans have different opinions on what makes a story creative. |
