Dataset · Biddls
The Onion Satirical News
Biddls/Onion_News
Corpus of satirical English news for headline generation, summarization, and text classification.
What's inside
The dataset contains 10K to 100K satirical articles from The Onion in English. The texts are presented as headline and article body pairs separated by a special #~# token. The repository also includes a script for independently parsing new issues.
What you can do with it
- Training models to generate humorous and satirical headlines
- Fine-tuning classifiers to distinguish satire and fakes from real news
- Testing models on text summarization tasks
- Tasks
- Summarization, Text generation, Text classification
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 182
- Likes
- 8
Models trained on it
- Wonder-Griffin/squadtrainTraXLText generation 163
- WonGrifferousAI/MisTraXLLMText generation 11
- llamadrama404/onion-fox-news-classifierText classification 11
- Wonder-Griffin/zeus-modelText generation 10
- Wonder-Griffin/ShorseyBeerLeagueText generation 6
- Wonder-Griffin/autotrain-blubawls1Text generation 0
- joshuachin/Qwen3-4B-Instruct-2507-satirical-headlines-CoTText generation 0
How to load it
from datasets import load_dataset
ds = load_dataset("Biddls/Onion_News")