Generating Data

---

LLMs have strong capabilities to generate coherent text. Using effective prompt strategies can steer the model to produce better, consistent, and more factual responses. LLMs can also be especially useful for generating data which is really useful to run all sorts of experiments and evaluations. For example, we can use it to generate quick samples for a sentiment classifier like so:

Prompt:

Output:

This is very useful. We actually use this example for a different test in another section of the guides.

Program-Aided Language ModelsGraduate Job Classification Case Study