metadata

license: mit
task_categories:
  - conversational
  - text-generation
language:
  - en
pretty_name: worldbuilding
size_categories:
  - 10K<n<100K

note: The dataset was fine, but the parquet bot appears to have messed it up somehow, if its not visible up there, look at dataset.jsonl

Updates:

Jan 7th 2024 - scraped all the worldbuilding stackexchange Q's, with 5+ rep, left with 18000 Q's.
Jan 8th 2024 - incorporated 100mb more in roleplay and worldbuilding, the dataset now includes Pippa, Bluemoon, and RolePlayIO
Jan 9th 2024 - More misc. worldbuilding data incorporated, the dataset is complete enough now

Worldbuild

A dataset focused on worldbuilding and roleplay, mostly well-formatted, high quality data, in the markdown format.