tool-calling

Definition

The capability where a model emits a structured call to an external function or API instead of an answer, then continues reasoning from whatever the call returns. It is what makes an LLM agent useful: search, arithmetic, database lookups, and code execution get delegated to systems that do them reliably. The term names the model-side behavior; the callable functions themselves are the tools.

Explainers using this term

• Paper Walkthrough: DART-SD — Training Tool-Calling Agents Without Flattening the DiamondDART-SD: Diamond-topology Aware Retrieval and Tuning for Self-Distillation of Multi-Turn Tool-Calling Agents