Dataset Viewer
Auto-converted to Parquet Duplicate
dataset_info
dict
sample_data
list
full_dataset_stats
dict
licensing
dict
{ "name": "HUBX Human Preference Dataset (PT-BR) - SAMPLE", "version": "1.0.0", "note": "This is a SAMPLE with 10 examples only. Contact for full dataset (450 tasks, 314k+ annotations).", "contact": "[email protected]" }
[ { "id": 1, "question": "Qual frase soa mais educada em um e-mail profissional?", "option_a": "Poderia, por gentileza, me enviar os dados até amanhã?", "option_b": "Preciso dos dados até amanhã.", "preferred": "A", "agreement_rate": 0.84, "total_votes": 699 }, { "id": 2, "question": "Qual resposta parece mais empática para alguém que perdeu o emprego?", "option_a": "Isso acontece com todo mundo, logo você arruma outro.", "option_b": "Sinto muito por isso. Se precisar de algo, estou por aqui.", "preferred": "B", "agreement_rate": 0.91, "total_votes": 702 }, { "id": 3, "question": "Qual descrição de produto é mais clara?", "option_a": "Um celular moderno com câmera boa.", "option_b": "Smartphone com câmera de 50MP, 128GB de armazenamento e bateria de longa duração.", "preferred": "B", "agreement_rate": 0.88, "total_votes": 685 }, { "id": 4, "question": "Qual título é mais atrativo para um artigo?", "option_a": "Como melhorar sua carreira.", "option_b": "5 passos comprovados para turbinar sua carreira em 2025.", "preferred": "B", "agreement_rate": 0.79, "total_votes": 710 }, { "id": 5, "question": "Qual resposta é mais útil para a pergunta 'Como economizar energia em casa?'", "option_a": "Tente apagar as luzes quando sair do cômodo.", "option_b": "Use lâmpadas LED, desconecte aparelhos da tomada e regule a temperatura do ar-condicionado.", "preferred": "B", "agreement_rate": 0.86, "total_votes": 698 }, { "id": 6, "question": "Qual frase comunica melhor empatia em uma conversa de suporte ao cliente?", "option_a": "Lamento pela sua experiência. Vamos resolver isso juntos.", "option_b": "Por favor, aguarde enquanto processamos sua solicitação.", "preferred": "A", "agreement_rate": 0.93, "total_votes": 715 }, { "id": 7, "question": "Qual frase é mais apropriada para um atendimento ao cliente?", "option_a": "Não sei como te ajudar.", "option_b": "Vou verificar isso para você e retorno em breve.", "preferred": "B", "agreement_rate": 0.96, "total_votes": 705 }, { "id": 8, "question": "Qual opção soa mais empática ao negar um pedido?", "option_a": "Infelizmente não será possível, mas estamos aqui para ajudar de outra forma!", "option_b": "Solicitação não autorizada. Procedimento negado.", "preferred": "A", "agreement_rate": 0.94, "total_votes": 688 }, { "id": 9, "question": "Qual mensagem é mais acolhedora para novo membro em grupo online?", "option_a": "Seja bem-vindo! Fique à vontade para interagir e crescer com a gente.", "option_b": "Membro adicionado. Utilize o grupo conforme as regras.", "preferred": "A", "agreement_rate": 0.89, "total_votes": 701 }, { "id": 10, "question": "Qual frase motiva mais alguém a continuar um curso online?", "option_a": "Você já chegou até aqui, falta pouco para concluir!", "option_b": "Módulo atual incompleto. Continue para finalizar.", "preferred": "A", "agreement_rate": 0.87, "total_votes": 695 } ]
{ "total_tasks": 450, "total_annotations": 314757, "unique_annotators": "~600", "avg_votes_per_task": 699, "language": "Portuguese (Brazil)", "period": "July - December 2024" }
{ "startup_license": "$4,500", "commercial_license": "$9,500", "contact": "[email protected]" }

🇧🇷 HUBX Human Preference Dataset (PT-BR)

The largest Portuguese-Brazilian human preference dataset for RLHF/DPO training.

📊 Dataset Statistics

Metric Value
Total Annotations 314,757
Unique Tasks 450
Human Annotators ~600
Avg. Votes per Task ~699
Language Portuguese (Brazil)
Domain Communication Quality & Tone

🎯 Why This Dataset?

  • 🇧🇷 Native PT-BR: Collected from Brazilian Portuguese speakers - not translated
  • 👥 Real Humans: ~600 verified annotators with payment records
  • 📈 High Agreement: ~70% inter-annotator agreement (proves human origin)
  • 💬 Justifications: Includes reasoning behind choices
  • 🎯 Practical Domain: Communication quality, empathy, tone, professionalism

🌎 Global Community

Our annotator community is expanding rapidly across multiple regions:

  • South America: Brazil, Argentina, Colombia, Chile, Peru, and more
  • North America: United States, Mexico
  • Europe: Portugal, Spain, Germany, UK
  • Asia: Japan, Philippines

This diversity enables us to create datasets in multiple languages and cultural contexts. Interested in custom datasets for other languages? Contact us.

📋 Use Cases

  • RLHF - Train reward models for Portuguese language models
  • DPO - Direct Preference Optimization for PT-BR
  • Tone Alignment - Make AI responses sound more natural and empathetic
  • Customer Service AI - Train chatbots for Brazilian market
  • UX Writing - Improve AI-generated microcopy

📁 Data Format

{
  "id": 1,
  "question": "Qual frase soa mais educada em um e-mail profissional?",
  "option_a": "Poderia, por gentileza, me enviar os dados até amanhã?",
  "option_b": "Preciso dos dados até amanhã.",
  "category": "professional_tone",
  "metadata": {
    "votes_a": 587,
    "votes_b": 112,
    "total_votes": 699,
    "preferred": "A",
    "agreement_rate": 0.84,
    "confidence": "high"
  },
  "justifications": [
    "A palavra 'por gentileza' demonstra mais respeito",
    "A opção A é mais cordial e profissional"
  ]
}

🔬 Data Collection

  • Platform: HUBX - Brazilian AI training community
  • Period: July - December 2024
  • Method: Paid annotators completing subscription-based tasks
  • Quality Control: Time-based filtering, justification collection
  • Verification: Payment records, unique user accounts, varied timestamps

Why This Proves Human Origin:

  1. Agreement Rate ~70% - AI would generate 95%+ agreement
  2. Justifications with typos - Natural Portuguese errors
  3. Payment Records - Real transactions for annotations
  4. ~600 Unique Accounts - Verified user registrations
  5. Temporal Distribution - Responses spread over time, not batch-generated

💰 Licensing

This dataset is available for commercial licensing.

📧 Contact for pricing: [email protected]

Pricing Tiers:

License Type Use Case Price
Research Academic/Non-commercial Free (with attribution)
Startup < $1M revenue $4,500
Commercial Full commercial rights $9,500
Enterprise Exclusive rights negotiable Contact

📬 Contact

📜 Citation

@dataset{hubx_ptbr_preferences_2024,
  title={HUBX Human Preference Dataset (PT-BR)},
  author={HUBX Community},
  year={2024},
  publisher={Hugging Face},
  language={Portuguese (Brazil)},
  size={314757 annotations}
}

⚡ Ready to improve your Portuguese AI? Contact us today.

Downloads last month
7