# Towards Building the Federated GPT: Federated Instruction Tuning

**Authors:** Jianyi Zhang, Saeed Vahidian, Martin Kuo, Chunyuan Li, Ruiyi Zhang, Tong Yu, Yufan Zhou, Guoyin Wang, Yiran Chen  
**arXiv:** [2305.05644](https://arxiv.org/abs/2305.05644)  
**Date:** 2023-05-09  
**Summary page:** https://www.summarizepaper.com/en/paper/2305.05644v2/

## Key points

- Authors introduce a novel approach called "Federated Instruction Tuning"
- Aim is to address challenges in acquiring high-quality instruction data for training large language models
- Existing practice of instruction-tuned models relies heavily on diverse and high-quality data like ChatGPT and GPT-4

## Summary

In their paper titled "Towards Building the Federated GPT: Federated Instruction Tuning," authors Jianyi Zhang, Saeed Vahidian, Martin Kuo, Chunyuan Li, Ruiyi Zhang, Tong Yu, Yufan Zhou, Guoyin Wang, and Yiran Chen introduce a novel approach called  to address the challenges associated with acquiring high-quality instruction data for training large language models (). The existing practice of instruction-tuned  relies heavily on diverse and high-quality data like ChatGPT and GPT-4.