unnatural_language / README.md
Melodyu's picture
Update README.md
82cb0da verified
metadata
task_categories:
  - text-classification
language:
  - en
tags:
  - climate
  - art
  - code
  - linguistics
  - ecology

The original data was collected by artist from the World Bank: https://documents.worldbank.org/en/publication/documents-reports

Public documents from the World Bank were cleaned and segmented into sentences.
Labeling was conducted on the Refinery platform using a hybrid, iterative human–machine workflow.

Here is the data artist provides:

  • WB_data_seqsonly.csv -> individual processed and cleaned sequences, gathered from many documents in one csv file
  • WB_dataset_labelled.csv -> sequences with labels (pro-growth / ecologically dominating / neutral); used in the current model
  • WB_dataset_separated_csv folder -> separated csv files, each file includes the cleaned sequences from the original document