Skip to content

Dataset · browndw

Parallel corpus of human and LLM texts in English

browndw/human-ai-parallel-corpus

A set of paired English text continuations from humans and neural networks for style study and AI generation detector training.

What's inside

Containing between 10 thousand and 100 thousand records, the dataset consists of English texts. To create it, a language model was given about 500 words of original text and asked to write another 500 words to compare the result with a real author's continuation of the same length.

What you can do with it

  • Train classifiers to recognize neural network-generated text
  • Comparative linguistic analysis of human and LLM writing styles
Tasks
Text classification
Languages
English
Size
10K–100K rows
Downloads this month
754
Likes
7

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("browndw/human-ai-parallel-corpus")