README.md

metadata

dataset_info:
  features:
    - name: category
      dtype: string
    - name: search_query
      dtype: string
    - name: search_type
      dtype: string
    - name: search_engine_input
      dtype: string
    - name: url
      dtype: string
    - name: title
      dtype: string
    - name: text_raw
      dtype: string
    - name: text_window
      dtype: string
    - name: stance
      dtype: string
  splits:
    - name: train
      num_bytes: 63378135
      num_examples: 4002
  download_size: 34731509
  dataset_size: 63378135
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
language:
  - en
license: mit
pretty_name: ConflictQA

This is an unofficial fork of the Conflict QA dataset introduced by Wan et al. (2024), please cite the original paper (citation below) if used.

dataset_info: features: - name: category dtype: string - name: search_query dtype: string - name: search_type dtype: string - name: search_engine_input dtype: string - name: url dtype: string - name: title dtype: string - name: text_raw dtype: string - name: text_window dtype: string - name: stance dtype: string splits: - name: train num_bytes: 63378135 num_examples: 4002 download_size: 34731509 dataset_size: 63378135 configs: - config_name: default data_files: - split: train path: data/train-* language: - en

Citation

@misc{wan2024evidence,
      title={What Evidence Do Language Models Find Convincing?}, 
      author={Alexander Wan and Eric Wallace and Dan Klein},
      year={2024},
      eprint={2402.11782},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}