Skip to content

Dataset · proj-persona

PersonaHub: Personas for Synthetic Data

proj-persona/PersonaHub

A collection of personas for generating diverse synthetic texts and instructions.

What's inside

This repository releases data from a paper on scaling synthetic data using 1 billion personas. It contains textual personas automatically gathered from the web in English and Chinese. The approach leverages diverse perspectives within LLMs to create varied synthetic data. The card specifies text-generation, classification, and other tasks.

What you can do with it

  • Generate synthetic SFT data using various personas
  • Train a classifier on persona descriptions
  • Evaluate the diversity approach in synthetic data
Tasks
Text generation, Text classification, Token classification, Fill mask, Table question answering
Languages
English, Chinese
Size
100K–1M rows
Downloads this month
5.8K
Likes
834

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("proj-persona/PersonaHub")