---
dataset_info:
- config_name: pair_feedback_iter_1
  features:
  - name: id
    dtype: string
  - name: query
    dtype: string
  - name: difficulty
    dtype: string
  - name: category
    dtype: string
  - name: quality
    dtype: string
  - name: intent
    dtype: string
  - name: source
    dtype: string
  - name: response
    dtype: string
  - name: feedback
    struct:
    - name: processed
      struct:
      - name: checklist
        dtype: string
      - name: intent
        dtype: string
      - name: score
        dtype: int64
      - name: strengths
        dtype: string
      - name: weaknesses
        dtype: string
    - name: raw
      dtype: string
    - name: model
      dtype: string
  - name: revision
    struct:
    - name: processed
      dtype: string
    - name: raw
      dtype: string
    - name: model
      dtype: string
  - name: pair_feedback
    struct:
    - name: model_A
      dtype: string
    - name: model_B
      dtype: string
    - name: processed
      struct:
      - name: analysis of A
        dtype: string
      - name: analysis of B
        dtype: string
      - name: choice
        dtype: string
      - name: reason of A=B
        dtype: string
      - name: reason of A>B
        dtype: string
      - name: reason of B>A
        dtype: string
    - name: raw_output
      dtype: string
    - name: revision_better
      dtype: bool
    - name: model
      dtype: string
  - name: model
    dtype: string
  splits:
  - name: train
    num_bytes: 550751588
    num_examples: 33116
  download_size: 268956837
  dataset_size: 550751588
- config_name: vapo
  features:
  - name: id
    dtype: string
  - name: query
    dtype: string
  - name: difficulty
    dtype: string
  - name: category
    dtype: string
  - name: quality
    dtype: string
  - name: intent
    dtype: string
  - name: source
    dtype: string
  splits:
  - name: train
    num_bytes: 60720299
    num_examples: 39349
  download_size: 33188220
  dataset_size: 60720299
configs:
- config_name: pair_feedback_iter_1
  data_files:
  - split: train
    path: pair_feedback_iter_1/train-*
- config_name: vapo
  data_files:
  - split: train
    path: vapo/train-*
---