Datasets:

gisub
/

cot_policy_word_ver1

dataset_info:
  features:
    - name: instruction
      dtype: string
    - name: input
      dtype: string
    - name: prompt
      dtype: string
    - name: output
      dtype: string
    - name: response
      dtype: string
    - name: cot_prompt
      dtype: string
  splits:
    - name: train
      num_bytes: 3200209
      num_examples: 362
  download_size: 880373
  dataset_size: 3200209
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*