Datasetpublic

hausmer/truha-news-satire

Truha news satire (Ukrainian) A synthetic, LLM-generated dataset of Ukrainian "news" written in the truha satirical register — short, meta-ironic, punchline-driven fake-news items. The corpus was produced by distilling a target style (a Ukrainian satirical news persona) into generated examples; no real user data is included. Intended use: style-transfer / imitation training for a Ukrainian satirical news-writing assistant. Each example is a (system, instruction, output) triple.… See the full description on the dataset page: https://huggingface.co/datasets/hausmer/truha-news-satire.

sourceHugging Faceotherupdated 12d agoView on Hugging Face

0likes47downloads

Dataset Card

Truha news satire (Ukrainian)

A synthetic, LLM-generated dataset of Ukrainian "news" written in the truha satirical register — short, meta-ironic, punchline-driven fake-news items. The corpus was produced by distilling a target style (a Ukrainian satirical news persona) into generated examples; no real user data is included.

Intended use: style-transfer / imitation training for a Ukrainian satirical news-writing assistant. Each example is a (system, instruction, output) triple.

Structure

Column	Type	Description
`system`	string	System prompt (persona: the satirical news voice).
`instruction`	string	The seed topic / hook the item should be written from.
`output`	string	The generated satirical news item (the target).
`style`	string	`truha_satire` (satirical news) or `general_ua` (general helper).

—train: 40,994 items — 39,272 truha_satire + 1,722 general_ua (general-helper event items; the same rows are published standalone in ukr-wiki-events).
—val: 2,066 held-out truha_satire items.

Notes

—Fully synthetic; no personally-identifiable information.
—Language: Ukrainian (mixed with the occasional Latin-script artifact, exactly as a real casual writer would type).
—For imitation training, feed system + instruction as the prompt and supervise on output.

License

Synthetic content. Licensed under the terms noted in the frontmatter. The generating and reference models were used to produce these examples.