---
dataset_info:
  features:
  - name: title
    dtype: string
  - name: abstract
    dtype: string
  - name: review
    dtype: string
  - name: source
    dtype: string
  - name: review_src
    dtype: string
  - name: thinking_trace
    dtype: string
  - name: conversations
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: review_sentences
    sequence: string
  - name: criteria_count
    struct:
    - name: criticism
      dtype: int64
    - name: example
      dtype: int64
    - name: importance_and_relevance
      dtype: int64
    - name: materials_and_methods
      dtype: int64
    - name: praise
      dtype: int64
    - name: presentation_and_reporting
      dtype: int64
    - name: results_and_discussion
      dtype: int64
    - name: suggestion_and_solution
      dtype: int64
    - name: total
      dtype: int64
  - name: criticism
    dtype: float64
  - name: example
    dtype: float64
  - name: importance_and_relevance
    dtype: float64
  - name: materials_and_methods
    dtype: float64
  - name: praise
    dtype: float64
  - name: presentation_and_reporting
    dtype: float64
  - name: results_and_discussion
    dtype: float64
  - name: suggestion_and_solution
    dtype: float64
  - name: total
    dtype: int64
  - name: has_criticism
    dtype: int64
  - name: has_example
    dtype: int64
  - name: has_importance_and_relevance
    dtype: int64
  - name: has_materials_and_methods
    dtype: int64
  - name: has_praise
    dtype: int64
  - name: has_presentation_and_reporting
    dtype: int64
  - name: has_results_and_discussion
    dtype: int64
  - name: has_suggestion_and_solution
    dtype: int64
  - name: reward_value
    dtype: float64
  - name: reward_value_length_adjusted
    dtype: float64
  - name: length_penalty
    dtype: float64
  splits:
  - name: train
    num_bytes: 19349983
    num_examples: 1729
  - name: val
    num_bytes: 198107442
    num_examples: 15096
  download_size: 83711889
  dataset_size: 217457425
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: val
    path: data/val-*
---