Skip to content

Dataset · Gryphe

English short stories by ChatGPT-4o

Gryphe/ChatGPT-4o-Writing-Prompts

3,746 cleaned stories based on Reddit Writing Prompts, generated by ChatGPT-4o, with titles and genre tags.

What's inside

The dataset includes 3,746 literary texts in English, each around 6,000 to 8,000 characters long. The stories were generated by the chatgpt-4o-latest model based on ideas from the Writing Prompts subreddit. The author cleaned the texts, preserving the original Markdown markup, and added a title and a list of suitable genres to each.

What you can do with it

  • SFT or LoRA training for long narrative story generation
  • Training literary genre classification and tagging models
Tasks
Text generation
Languages
English
Size
1K–10K rows
Downloads this month
4.6K
Likes
38

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("Gryphe/ChatGPT-4o-Writing-Prompts")