Loading Now

Summary of On Training Data Influence Of Gpt Models, by Yekun Chai et al.


On Training Data Influence of GPT Models

by Yekun Chai, Qingyi Liu, Shuohuan Wang, Yu Sun, Qiwei Peng, Hua Wu

First submitted to arxiv on: 11 Apr 2024

Categories

  • Main: Computation and Language (cs.CL)
  • Secondary: Machine Learning (cs.LG)

     Abstract of paper      PDF of paper


GrooveSquid.com Paper Summaries

GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!

Summary difficulty Written by Summary
High Paper authors High Difficulty Summary
Read the original abstract here
Medium GrooveSquid.com (original content) Medium Difficulty Summary
The proposed approach, GPTfluence, investigates how training data affects the performance of Generative Pre-trained Transformer (GPT) models. It uses a featurized simulation to track the influence of individual training instances on loss and other key metrics at targeted test points. This method enables comparison with existing approaches across various scenarios and tasks, including natural language understanding and generation. The results demonstrate robust generalization capabilities to unseen training data in both fine-tuning and instruction-tuning settings. The code and data are publicly available.
Low GrooveSquid.com (original content) Low Difficulty Summary
GPTfluence is a new way to study how the data used to train GPT models affects their performance. Researchers created a special simulation that shows how individual pieces of training data affect the model’s behavior. This helps us understand how well the model will work on new, unseen data. The results are promising and show that GPTfluence can be used in different settings, like fine-tuning or instruction-tuning, to make language models more accurate.

Keywords

» Artificial intelligence  » Fine tuning  » Generalization  » Gpt  » Instruction tuning  » Language understanding  » Transformer