CoolFace
Datasetpublic

Largefilly/HEAR-Hispanic_Emotional_Accompaniment_Responses

HEAR Dataset Description The HEAR (Hispanic Emotional Accompaniment Responses) dataset is designed to train language models in the task of emotionally accompanying users. This dataset enables models to generate empathetic and appropriate responses in Spanish, understanding and responding to different emotional situations. Dataset Origin The HEAR dataset was created using elements from the HRECPW dataset, which contains 11 emotional categories with… See the full description on the dataset page: https://huggingface.co/datasets/Largefilly/HEAR-Hispanic_Emotional_Accompaniment_Responses.

sourceHugging Facemitupdated 5mo agoView on Hugging Face
0likes14downloads
Dataset Card

HEAR Dataset

Description

The HEAR (Hispanic Emotional Accompaniment Responses) dataset is designed to train language models in the task of emotionally accompanying users. This dataset enables models to generate empathetic and appropriate responses in Spanish, understanding and responding to different emotional situations.

Dataset Origin

The HEAR dataset was created using elements from the HRECPW dataset, which contains 11 emotional categories with 11,000 samples each. Synthetic data generated with OpenAI's GPT-3.5 model was used to balance the dataset, ensuring that each emotional category was equally represented.

Dataset Structure

The HEAR dataset is composed of three subsets: training, validation, and test. Each of these subsets contains balanced samples for the 11 emotional categories, distributed as follows:

  • —Training: 41,910 samples
  • —Validation: 2,200 samples (200 per category)
  • —Test: 1,320 samples (120 per category)

Emotional Categories

The dataset covers the following emotional categories:

  1. 1.Affection
  2. 2.Joy
  3. 3.Admiration
  4. 4.Anger
  5. 5.Sadness
  6. 6.Optimism
  7. 7.Hate
  8. 8.Surprise
  9. 9.Fear
  10. 10.Calm
  11. 11.Disgust

Data Generation Process

To create synthetic responses, OpenAI's GPT-3.5 model was used via its API. Specific prompts were designed to generate responses in Spanish that were empathetic and contextually appropriate to the user's emotional message. An example of a prompt used is as follows:

markdown
Below is an instruction that describes a task, paired with an input that provides further context.

### instruction:
 You are an emotional assistant, respond empathetically in Spanish to each of the messages. If the user seems sad or upset, you should offer words of encouragement. If the user seems happy or excited, the assistant should share that joy and respond enthusiastically. In all cases, the assistant should maintain a respectful tone, if possible, encouraging you to talk more about it. Don't say hello, unless necessary. Use the username and pronoun to respond in a personalized way.

### input:
{"Username":"Sohel", "pronoun":"el", "message":"El miedo me paraliza cuando tengo que tomar decisiones importantes en mi vida."}

### response:
Entiendo que el miedo te paralice al tomar decisiones importantes en tu vida. Recuerda que tomar decisiones es parte natural del crecimiento y que siempre hay opciones y apoyo disponible si te sientes abrumado. ¿Hay alguna decisión específica que te preocupe en este momento?</s>

License

Sólo Escúchame is released under the MIT License.

Citation

If you use HEAR in your research, please cite the following paper:

bibtex
@misc{ramírez2024soloescuchamespanishemotional,
      title={S\'olo Esc\'uchame: Spanish Emotional Accompaniment Chatbot}, 
      author={Bruno Gil Ramírez and Jessica López Espejel and María del Carmen Santiago Díaz and Gustavo Trinidad Rubín Linares},
      year={2024},
      eprint={2408.01852},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2408.01852}, 
}

Contact

For any questions or inquiries, please contact:

  • —Bruno Gil Ramírez: bruno.gilram@gmail.com