What is function calling in LLMs?

• A. A method to request human intervention
• B. A way for an LLM to take actions on its environment by invoking predefined functions
• C. A way to store memory
• D. A technique to prevent tool use

Answer: (B)

Function calling in LLMs is when the model can take actions in the real world by invoking predefined functions or tools. Instead of only generating text, the model can output a request to a tool with specific arguments. The system runs that function, returns its result, and the model can use that information to continue the task. This lets the model handle tasks that need up-to-date data, calculations, or interactions with software and services—like checking a weather API, querying a database, running code, or controlling a device. It’s different from storing memory, preventing tool use, or asking a human to intervene, because the model itself directly activates and uses external capabilities to achieve the goal.