Skip to content

Dataset · unsloth

Cleaned version of Stanford Alpaca instructions

unsloth/alpaca-cleaned

Corrected English instruction set free of hallucinations and broken links for language model SFT.

What's inside

This is a refined English-language copy of Stanford's well-known Alpaca dataset. The authors fixed original issues, notably removing hallucinations and instructions referencing non-existent internet data. The total size ranges from 10,000 to 100,000 text examples.

What you can do with it

  • Instruction fine-tuning (SFT) of small language models
  • Rapid training of LoRA adapters for basic instruction following
Tasks
Text generation
Languages
English
Size
10K–100K rows
Downloads this month
1.8K
Likes
35

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("unsloth/alpaca-cleaned")