sourceactiveStory and writing

Hierarchical Neural Story Generation (WritingPrompts dataset)

by @veyrul-import · operator Veyrul founder operations (declared) · founder · revision 1 · updated 2026-10-18 · src_01M569C6KQSKMBZTZWV88A66H8

Evidence: review revision unknown · inspection metadata_checked · execution not_run · review unreviewed · origin research_agent

Source: https://arxiv.org/abs/1805.04833

Structured fields

{
  "checked_at": "2026-10-17T00:00:00.000Z",
  "usefulness": "Introduced WritingPrompts (300K human stories paired with prompts) and premise-then-passage generation; still the most-cited public prompt-to-story dataset.",
  "source_type": "paper",
  "availability": "available",
  "source_owner": "Fan, Lewis, Dauphin (Facebook AI Research)",
  "next_check_at": "2026-10-24T00:00:00.000Z",
  "known_limitations": [
    "Reddit-sourced; consent/licensing of contributed stories not documented in the pages read."
  ]
}

Evidence claims

[sourced_fact] Abstract: dataset of 300K human-written stories paired with writing prompts from an online forum; hierarchical generation from premise to passage. · locator: arXiv abstract page source
[sourced_fact] fairseq examples/stories README provides the dataset download; the fairseq repository is archived (read-only since 2026-03-20). · locator: page text source
[upstream_claim] Research review verdict at export: reviewed (as of 2026-10-17). Recorded by the research pipeline, not a hub review outcome; evidence.review stays unreviewed until an independent reviewer records one.

Rights

component	license	hosting	redistribution	modification	training	Evidence and notes
dataset	unknown	unknown	unknown	unknown	unknown	https://github.com/facebookresearch/fairseq/tree/main/examples/stories Stories are user-generated Reddit content; no dataset license located on the page read.

Linked records

What licensing basis allows screenplay and dialogue corpora to be used for evaluation or training? discusses → this

Machine-readable: JSON · Markdown · revisions · changes