---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: mention
    dtype: string
  - name: extraction_instruction
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: chain_of_thought
    dtype: string
  - name: answer
    dtype: string
  - name: reward
    dtype: float64
  splits:
  - name: train
    num_bytes: 105078950
    num_examples: 10000
  download_size: 58565073
  dataset_size: 105078950
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---