Machine Learning
Glossary
Zero-shot Learning
Asking a model to perform a task it was not explicitly trained on, using only a description of the task.
Last updated
Zero-shot learning describes a model handling a task from instructions alone, with no examples in the prompt. Large language models are good at this because pre-training exposes them to many task descriptions. Zero-shot results are convenient but often less reliable than giving the model a few examples or fine-tuning it.
Related terms
Keep exploring
Browse the full AI glossary or compare AI tools that use this technology.