HatEval - Multilingual Detection of Hate Speech Against Immigrants and Women

# HatEval - Multilingual Detection of Hate Speech
# Based on Basile et al., SemEval 2019
# Paper: https://aclanthology.org/S19-2007/
# Dataset: https://competitions.codalab.org/competitions/19935
#
# This task asks annotators to first determine whether a tweet contains
# hate speech, and if so, to identify the types of hate expressed.
#
# Hate Speech Labels:
# - Hateful: The tweet contains hate speech targeting a group
# - Not Hateful: The tweet does not contain hate speech
#
# Hate Type Categories:
# - Racism: Targeting based on race or ethnicity
# - Sexism: Targeting based on gender
# - Xenophobia: Targeting based on national origin or immigration status
# - Other: Other forms of hate not covered above

annotation_task_name: "HatEval - Hate Speech Detection"
task_dir: "."

data_files:
  - sample-data.json

item_properties:
  id_key: "id"
  text_key: "text"

output_annotation_dir: "annotation_output/"
output_annotation_format: "json"

port: 8000
server_name: localhost

annotation_schemes:
  - annotation_type: radio
    name: hate_speech_detection
    description: "Does this tweet contain hate speech?"
    labels:
      - "Hateful"
      - "Not Hateful"
    keyboard_shortcuts:
      "Hateful": "1"
      "Not Hateful": "2"
    tooltips:
      "Hateful": "The tweet contains language that targets or demeans a group"
      "Not Hateful": "The tweet does not contain hate speech"

  - annotation_type: multiselect
    name: hate_type
    description: "If hateful, what types of hate are expressed? Select all that apply."
    labels:
      - "Racism"
      - "Sexism"
      - "Xenophobia"
      - "Other"
    tooltips:
      "Racism": "Hate speech targeting based on race or ethnicity"
      "Sexism": "Hate speech targeting based on gender"
      "Xenophobia": "Hate speech targeting based on national origin or immigration status"
      "Other": "Other forms of hate not covered by the above categories"

annotation_instructions: |
  You will be shown a tweet and its target group. Your task is to:
  1. Determine whether the tweet contains hate speech (hateful vs. not hateful).
  2. If hateful, select all applicable hate type categories.
  Consider the context and target group when making your judgment.

html_layout: |
  <div style="padding: 15px; max-width: 800px; margin: auto;">
    <div style="background: #fefce8; border: 1px solid #fde68a; border-radius: 8px; padding: 12px; margin-bottom: 12px;">
      <strong style="color: #a16207;">Target Group:</strong>
      <span style="font-size: 15px;">{{target_group}}</span>
    </div>
    <div style="background: #f0f9ff; border: 1px solid #bae6fd; border-radius: 8px; padding: 16px; margin-bottom: 16px;">
      <strong style="color: #0369a1;">Tweet:</strong>
      <p style="font-size: 16px; line-height: 1.7; margin: 8px 0 0 0;">{{text}}</p>
    </div>
  </div>

allow_all_users: true
instances_per_annotator: 50
annotation_per_instance: 2
allow_skip: true
skip_reason_required: false