Skip to main content
Datasetinstruction-tuningv2.0

Dolly-15K

by Databricks · open-source · Last verified 2026-03-17

A high-quality, human-generated instruction dataset created by Databricks employees, containing 15,000 prompt/response pairs across 8 capability categories. Designed to demonstrate that instruction-following capability can be instilled via a modest amount of human-written data.

https://huggingface.co/datasets/databricks/databricks-dolly-15k
B
BAbove Average
Adoption: AQuality: B+Freshness: BCitations: AEngagement: F

Specifications

License
CC-BY-SA-3.0
Pricing
open-source
Capabilities
instruction-tuning, supervised-fine-tuning
Integrations
huggingface-datasets
Use Cases
fine-tuning, instruction-following, chatbot-training
API Available
No
Tags
instruction-following, human-generated, databricks, fine-tuning
Added
2026-03-17
Completeness
100%

Index Score

68.3
Adoption
80
Quality
79
Freshness
68
Citations
82
Engagement
0

Put AI to work for your business

Deploy this dataset alongside autonomous AaaS agents that handle tasks end-to-end — no babysitting required.

Explore the full AI ecosystem on Agents as a Service