mchl-labs/stambecco_data_plus_it

🌁 Stambecco-Plus: Premium Italian Instruction-Tuning Dataset

💡 Note: This is the premium GPT-4 distilled version. If you are looking for the original Stambecco dataset based on Alpaca-Cleaned, you can find it here.

---

📌 Dataset Summary

This repository contains the dataset used to train the Stambecco Plus models.

The dataset is a high-quality Italian translation and adaptation of the Alpaca-GPT4 dataset. Because the underlying responses were distilled from GPT-4 rather than GPT-3.5, this dataset features superior reasoning, richer vocabulary, and far fewer hallucinations than standard Alpaca datasets.

• Language: Italian (it)
• Base Source: Alpaca-GPT4 (Instruction Tuning with GPT-4)
• Primary Use Case: High-quality instruction fine-tuning, evaluation, and alignment for Italian LLMs.
• Repository: mchl-labs/stambecco_data_plus_it

---

📖 Data Structure

Each entry follows the standard Alpaca instruction-tuning format:

{
  "instruction": "Spiega i vantaggi dell'energia solare rispetto ai combustibili fossili.",
  "input": "",
  "output": "L'energia solare offre numerosi vantaggi rispetto ai combustibili fossili. In primo luogo, è una fonte rinnovabile e inesauribile..."
}

• instruction: The prompt or task description in Italian.
• input: Optional contextual information required for the task.
• output: The target Italian response.

---

🚀 Quickstart

Load the dataset directly using the Hugging Face datasets library:

from datasets import load_dataset

dataset = load_dataset("mchl-labs/stambecco_data_plus_it")
print(dataset["train"][0])

---

📜 Citation & Attribution

If you use this dataset in your research, please include it in your paper's formal bibliography/references section (rather than a footnote) so citation engines can properly track attribution.

1. Cite Stambecco-Plus (The Italian Dataset)

@misc{stambecco_plus_2023,
  author       = {Michael Rottoli},
  title        = {Stambecco-Plus: Premium Italian Instruction-Tuning Dataset},
  year         = {2023}, 
  publisher    = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/mchl-labs/stambecco_data_plus_it}}
}

2. Original Source Attribution

This dataset is a translated derivative of the Alpaca GPT-4 dataset and the original Stanford Alpaca architecture. If you cite Stambecco Plus, please also consider citing the original researchers:

Alpaca-GPT4:

@article{peng2023gpt4llm,
    title={Instruction Tuning with GPT-4},
    author={Baolin Peng and Chunyuan Li and Pengcheng He and Michel Galley and Jianfeng Gao},
    journal={arXiv preprint arXiv:2304.03277},
    year={2023}
}

Stanford Alpaca (Original base dataset):

@misc{alpaca,
  author       = {Rohan Taori and Ishaan Gulrajani and Tianyi Zhang and Yann Dubois and Xuechen Li and Carlos Guestrin and Percy Liang and Tatsunori B. Hashimoto},
  title        = {Stanford Alpaca: An Instruction-following LLaMA model},
  year         = {2023},
  publisher    = {GitHub},
  howpublished = {\url{https://github.com/tatsu-lab/stanford_alpaca}}
}

---

⚖️ License

This dataset is released under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license.

It is provided strictly for non-commercial academic research. Because the underlying data was generated using OpenAI's GPT-4 API, this dataset is subject to OpenAI's Terms of Use and cannot be used to develop models that compete commercially with OpenAI. If you remix, transform, or build upon this dataset, you must distribute your contributions under the same open, non-commercial license.