| --- |
| license: mit |
| tags: |
| - text-generation |
| - instruction-tuning |
| - tool-calling |
| - reasoning |
| - brazilian-knowledge |
| - bilingual |
| - uirapuru |
| - synastria |
| pretty_name: Uirapuru U1.1 Bilingual Reasoning & Tool Calling Dataset |
| size_categories: |
| - 1K |
| task_categories: |
| - text-generation |
| language: |
| - pt |
| - en |
| --- |
| |
| # Uirapuru U1.1 Bilingual Reasoning & Tool Calling Dataset |
|
|
| ### Summary |
| Uirapuru U1.1 is a specialized bilingual dataset designed to enhance Large Language Models (LLMs) in three critical areas: **Tool Calling**, **Brazilian Local Knowledge**, and **Human-like Reasoning**. Created by SynastrIA Networks, this dataset bridges the gap between generic multilingual models and agents capable of operating effectively within the Brazilian context while maintaining strong alignment with English-centric benchmarks. |
|
|
| ### Key Features |
| - **🛠️ Advanced Tool Calling:** Structured examples for function calling, API interaction, and agentic workflows in both PT-BR and EN-US. |
| - **🇷 Deep Brazilian Knowledge:** Curated content covering Brazilian geography, culture, legislation, slang, and local nuances often missing in global datasets. |
| - **🧠 Human-Centric Reasoning:** Conversational chains-of-thought that mimic natural human problem-solving patterns rather than rigid logical templates. |
| - **⚖️ Balanced Bilingualism:** High-quality parallel and monolingual samples ensuring equal proficiency in both languages. |
|
|
| ## Uses |
|
|
| ### Direct Use |
| This dataset is optimized for: |
| - Fine-tuning LLMs for **agentic tasks** and tool use in Brazil. |
| - Improving **cultural alignment** and reducing hallucinations about Brazilian topics. |
| - Enhancing **reasoning capabilities** to sound more natural and less robotic. |
| - Cross-lingual transfer of tool-calling abilities from English to Portuguese. |
|
|
| ### Out-of-Scope Use |
| Not intended for general pre-training from scratch due to its specialized nature. Should not be used for tasks requiring verified legal or medical advice without human oversight. |
|
|
| ## Dataset Info. |
|
|
| ### Curation Rationale |
| Most existing datasets either lack high-quality Portuguese tool-calling examples or fail to capture the "human touch" in reasoning. Uirapuru U1.1 was built to create agents that don't just translate English behaviors but truly understand and operate within the Brazilian ecosystem with natural intelligence. |
|
|
| ### Annotations |
| Data was validated by native Brazilian Portuguese speakers and English experts to ensure cultural accuracy and reasoning quality. Tool-calling formats were verified against common API standards. |
|
|
| ## Additional Information |
|
|
| ### Licensing Information |
| MIT License. Free for commercial and research use. |
|
|
| ### Citation |
| If you use Uirapuru U1.1 in your work, please cite: |
| ```bibtex |
| @misc{uirapuru_u1_1, |
| author = {SynastrIA Networks}, |
| title = {Uirapuru U1.1: Bilingual Tool Calling, Brazilian Knowledge & Human Reasoning Dataset}, |
| year = {2024}, |
| publisher = {Hugging Face}, |
| url = {https://huggingface.co/datasets/gustajunq/uirapuru-dataset} |
| } |
| ``` |
|
|
| ### Contact |
| For questions, collaborations, or issues, please contact SynastrIA Networks or open an issue on this repository. |
