OpenAI Developer Forum
For which classification tasks, zero-shot GPT-3 would outperform fine-tuned Bert?
API
MichaelMarkevych
July 15, 2022, 6:32pm
1
It’s rather a theoretical question. Suppose we have either 1,000 or 10,000 examples for fine-tuning.
Related Topics
Topic
Replies
Views
Activity
Accuracy of few-shot learning vs. fine-tuning for tens of examples
API
11
5132
February 20, 2024
Zero Shot Learning on GPT 4 vs. Fine Tuning 3.5 for chat?
API
4
1399
July 7, 2023
Help with fine-tuning for text categorization
API
4
773
December 16, 2023
Best way to get GPT to output a DSL
API
api
0
393
December 8, 2023
Fine-Tuning Classifiers; Poor Performance
API
api
5
612
August 27, 2023