port: 8080
annotation_task_name: LLM Response Quality Comparison
task_dir: .
output_annotation_dir: annotation_output/llm-preference/
output_annotation_format: json
data_files:
- data/llm_responses.json
item_properties:
  id_key: id
  text_key: responses
list_as_text:
  text_list_prefix_type: alphabet
user_config:
  allow_all_users: true
  users: []
annotation_schemes:
- annotation_type: pairwise
  name: overall_preference
  description: Overall, which response is better?
  mode: scale
  items_key: responses
  labels:
  - Response A
  - Response B
  scale:
    min: 1
    max: 6
    step: 1
    default: 3
    labels:
      min: A is much better
      max: B is much better
      center: ''
  label_requirement:
    required: true
  layout:
    columns: 1
- annotation_type: pairwise
  name: helpfulness
  description: Which response is more helpful for the user's actual need?
  mode: binary
  items_key: responses
  labels:
  - A
  - B
  allow_tie: true
  tie_label: Equally helpful
  sequential_key_binding: true
  layout:
    columns: 1
- annotation_type: pairwise
  name: tone
  description: Which response has a more appropriate tone?
  mode: binary
  items_key: responses
  labels:
  - A
  - B
  allow_tie: true
  tie_label: Both appropriate
  sequential_key_binding: false
  layout:
    columns: 1
- annotation_type: text
  name: justification
  description: Briefly explain your preference (optional)
  label_requirement:
    required: false
  layout:
    columns: 2
site_dir: default
require_no_password: true