---
dataset_info:
  features:
  - name: question
    dtype: string
  - name: answer
    dtype: string
  splits:
  - name: train
    num_bytes: 210193
    num_examples: 200
  download_size: 147910
  dataset_size: 210193
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
task_categories:
- text2text-generation
language:
- en
pretty_name: Discord Conversations
size_categories:
- 1K<n<10K
---
A question-answer dataset of real messages from Discord, all filtered to avoid accidentally doxxing anyone.