Agents

In-Context Learning

1 min read

What It Means

The ability of LLMs to learn from examples provided in the prompt without updating model weights.

In-context learning (ICL) enables rapid adaptation through few-shot examples, a key capability for versatile agents.

How It Works

Examples provided in prompt
Model infers pattern
Applies to new inputs
No training required

Limitations

Context window constraints
Not persistent
Variable reliability

Explore evaluation patterns

See how this applies in practice

agentscapabilitieslearning

Back to Glossary