---
dataset_info:
  features:
  - name: dataset
    dtype: string
  - name: generator
    dtype: string
  - name: instruction
    dtype: string
  - name: output
    dtype: 'null'
  - name: messages
    list:
    - name: role
      dtype: string
    - name: content
      dtype: string
  - name: compare_id
    dtype: string
  splits:
  - name: train
    num_bytes: 265352
    num_examples: 2000
  download_size: 71469
  dataset_size: 265352
configs:
- config_name: alpaca_eval_gpt4_baseline
  data_files:
  - split: eval
    path: data/train-*
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---