Advanced techniques
In previous sections, we covered some basics techniques of prompt engineering. Those techniques should be kept in mind regardless of the type of application your are developing, since are general best practices that improve your LLM performance anyway.On the other hand, there are some advanced techniques which might be implemented for specific scenarios, that we are going to cover in the upcoming sections.
Few-shot approach
In their paper “Language Models are Few-Shot Learners”, the authors demonstrate that GPT-3 can achieve strong performance on many NLP tasks in a few-shot setting. This means that for all tasks, GPT-3 is applied without any fine-tuning, with tasks and few-shot demonstrations specified purely via text interaction with the model.This is an example and evidence of how the concept of few-shot learning – which means, providing the models with examples of how we would like it to respond – is a powerful techniques that...
