Dataset · Hugging Face

CohereLabs/aya_dataset

Name: aya_dataset
Creator: CohereLabs
Keywords: task_categories:other, annotations_creators:crowdsourced, annotations_creators:expert-generated, language_creators:crowdsourced, language_creators:expert-generated, multilinguality:multilingual, source_datasets:original, language:amh, language:arb, language:ary, language:ars, language:acq, language:arz, language:apc, language:ben

Dataset Summary The Aya Dataset is a multilingual instruction fine-tuning dataset curated by an open-science community via Aya Annotation Platform from Cohere Labs. 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 &… See the full description on the dataset page: https://huggingface.co/datasets/CohereLabs/aya_dataset.

View on Hugging FaceReport an issue

Author / organization	CohereLabs
Downloads	23.2K (as reported by Hugging Face)
Likes	377
Created	31 January 2024
Last updated at source	15 April 2025

Embed this entry

<iframe src="https://silytics.si/embed/item/1ae6b42b-281c-464b-b5f5-edfe328f6f3f" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Connected in the knowledge graph

ig1/Qwen3.5-9B-NVFP4Model · used to train

Explore this entry in the graph

More from CohereLabs

CohereLabs/xP3xDataset
CohereLabs/Global-MMLUDataset
CohereLabs/aya_collectionDataset
CohereLabs/Global-MMLU-LiteDataset
CohereLabs/aya_collection_language_splitDataset

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 15 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.