Zero-Shot Learning

Level 1

When the AI can do something it's never been shown how to do — like guessing what a new animal is just from its description.

Level 2

When AI can perform a task without any examples — you just describe what you want and it figures it out. This is one of the most impressive things about modern LLMs.

Level 3

A model's ability to perform tasks it wasn't explicitly trained on, using only a natural language description of the task without any demonstration examples.

Level 4

Task performance without any in-context demonstrations, relying entirely on the model's pre-trained knowledge and instruction-following capabilities to generalize from the task description alone.

Level 5

Generalization to novel task specifications at inference time without parameter updates or in-context exemplars — enabled by instruction tuning and the model's implicit task distribution learned during pre-training, analyzable through the lens of meta-learning and compositional generalization.

Sources

• NIST Trustworthy and Responsible AI Resource Center: Glossary — National Institute of Standards and Technology. Accessed 2026-07-20.