---
dataset_info:
  features:
  - name: action_uid
    dtype: string
  - name: operation
    dtype: string
  - name: pos_candidates
    sequence: string
  - name: neg_candidates
    sequence: string
  - name: website
    dtype: string
  - name: domain
    dtype: string
  - name: subdomain
    dtype: string
  - name: annotation_id
    dtype: string
  - name: confirmed_task
    dtype: string
  - name: action_reprs
    sequence: string
  - name: target_action_index
    dtype: string
  - name: target_action_reprs
    dtype: string
  - name: action
    dtype: string
  - name: original_action_repr
    dtype: string
  - name: original_pos_candidate
    struct:
    - name: attributes
      struct:
      - name: alt
        dtype: string
      - name: aria_description
        dtype: string
      - name: aria_label
        dtype: string
      - name: backend_node_id
        dtype: string
      - name: bounding_box_rect
        dtype: string
      - name: class
        dtype: string
      - name: data_pw_testid_buckeye_candidate
        dtype: string
      - name: id
        dtype: string
      - name: input_checked
        dtype: string
      - name: input_value
        dtype: string
      - name: is_clickable
        dtype: string
      - name: label
        dtype: string
      - name: name
        dtype: string
      - name: placeholder
        dtype: string
      - name: role
        dtype: string
      - name: text_value
        dtype: string
      - name: title
        dtype: string
      - name: type
        dtype: string
      - name: value
        dtype: string
    - name: backend_node_id
      dtype: string
    - name: is_original_target
      dtype: bool
    - name: is_top_level_target
      dtype: bool
    - name: tag
      dtype: string
  - name: match_type
    dtype: string
  - name: cleaned_accessibility_tree
    dtype: string
  - name: previous_actions
    sequence: string
  - name: cleaned_next_accessibility_tree
    dtype: string
  - name: next_state_tao
    dtype: string
  - name: new_items
    dtype: string
  - name: updated_items
    dtype: string
  - name: deleted_items
    dtype: string
  - name: refined_tao
    dtype: string
  - name: raw_prediction
    dtype: string
  - name: rationale
    dtype: string
  - name: next_state_description_with_tao
    dtype: string
  - name: reward_score
    dtype: string
  splits:
  - name: train
    num_bytes: 1367938090
    num_examples: 6125
  download_size: 274000502
  dataset_size: 1367938090
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---