Mythos
收藏资源简介:
Mythos数据集是由麻省大学阿默斯特分校和马里兰大学帕克分校构建的,包含64位作者创作的590个故事,这些故事来源于Reddit、AO3、Storium、《叙事杂志》和《纽约客》等五个不同的平台。该数据集旨在支持作者风格分析和个人化故事生成评价,通过将每个作者的作品按时间顺序分为两部分,一部分用于推断写作特征,另一部分用于评估LLM模仿作者独特写作行为的能力。
The Mythos Dataset was constructed by the University of Massachusetts Amherst and the University of Maryland, College Park. It contains 590 stories created by 64 authors, sourced from five distinct platforms: Reddit, AO3, Storium, Narrative Magazine, and The New Yorker. This dataset aims to support author style analysis and personalized story generation evaluation. Specifically, the works of each author are split into two chronological subsets: one for inferring writing characteristics, and the other for evaluating the ability of LLMs to mimic the author's unique writing behaviors.

- 1Whose story is it? Personalizing story generation by inferring author styles麻省大学阿默斯特分校,马里兰大学帕克分校 · 2025年



