LLM-2307.04346.pdf

1. The paper proposes using Large Language Models (LLMs) to generate property-based tests for API documentation.
2. The LLMs can write logic for generating random inputs and derive meaningful property assertions from the API documentation.
3. The potential of using LLMs for Property-Based Testing (PBT) is demonstrated in Figure 2.
4. The LLMs can be instructed to write generator functions or property assertions to check the properties of the result.
5. The study investigates whether LLMs can write good property-based tests for APIs.