Datasets:

jason-oneal
/

dnd-5e-dataset

It follows the Alpaca format — each example is an instruction, an optional input, and an output.
This dataset helps train models that can answer questions, generate content, and assist as an AI Dungeon Master or lore assistant.

✅ Content

This dataset covers:

Ability scores
Alignments
Backgrounds & features
Classes & mechanics
Conditions & effects
Damage types
Equipment, weapons & armor
Feats & features
Languages
Magic items & schools
Monsters & stat blocks
Proficiencies
Races & subraces
Rules & rule sections
Skills
Spells & descriptions
Subclasses
Traits
Weapon properties

📂 Structure

Each data point includes:

instruction: The prompt or question.
input: Additional context (can be empty).
output: The answer or response.

Example:

{
  "instruction": "What is the ability score 'Charisma' in D&D 5e?",
  "input": "",
  "output": "Charisma (CHA) measures your ability to interact effectively with others. It includes such factors as confidence and eloquence, and it can represent a charming or commanding personality. A Charisma check might arise when you try to influence or entertain others, when you try to make an impression or tell a convincing lie, or when you are navigating a tricky social situation. The Deception, Intimidation, Performance, and Persuasion skills reflect aptitude in certain kinds of Charisma checks."
}

🔗 Usage

Load with the Hugging Face datasets library:

from datasets import load_dataset

dataset = load_dataset("jason-oneal/dnd-5e-dataset")
print(dataset["train"][0])

🚀 Intended Uses

Train custom chatbots for D&D 5e knowledge.

Build an AI Dungeon Master.
Generate monster stat blocks, magic items, or NPC lore.
Clarify D&D 5e rules with an assistant.

📜 Citation

If you use this dataset, please cite it:

@misc{dnd5e-alpaca,
  author = {jason-oneal},
  title = {D&D 5e Alpaca Dataset},
  year = {2025},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/jason-oneal/dnd-5e-dataset}}
}

⚙️ License

MIT License — you are free to use, modify, and build on this dataset.