---
language:
- pt
pretty_name: Aya Dataset Portuguese
tags:
- aya
- portuguese
- legal
- chemistry
license: apache-2.0
size_categories:
- 1K<n<10K
---

CohereForAI [Aya Dataset](https://huggingface.co/datasets/CohereForAI/aya_dataset) filtrado para português (PT).

**Aya Dataset Summary**

The [Aya Dataset](https://huggingface.co/datasets/CohereForAI/aya_dataset) is a multilingual instruction fine-tuning dataset curated by an open-science community via Aya Annotation Platform from Cohere For AI. The dataset contains a total of 204k human-annotated prompt-completion pairs along with the demographics data of the annotators.
This dataset can be used to train, finetune, and evaluate multilingual LLMs.

Curated by: Contributors of Aya Open Science Intiative.

Language(s): 65 languages (71 including dialects & scripts).

License: Apache 2.0