Questions
Common questions.
What is Zero-Shot Learning?
Zero-shot learning is when a model performs a task it was not explicitly trained or given examples for, relying solely on its pre-trained knowledge and the task instruction.
How does Zero-Shot Learning work?
Agent governance must account for the higher unpredictability of zero-shot performance.
Which terms are related to Zero-Shot Learning?
Closely related concepts include Model Factsheet, Few-Shot Learning, Foundation Model, System Prompt. Each is defined in the Prefactor glossary.
Keep reading