Zero-Shot Learning.
Learn what Zero-Shot Learning means in modern search and SEO.
An AI capability where a model performs a task it has never been explicitly trained on, using only its general pre-trained knowledge.
Zero-shot learning refers to a model's ability to handle tasks it wasn't specifically trained for, by drawing on its broad pre-trained knowledge. If you ask an LLM to classify a product review as positive or negative without providing any examples, and it does so correctly, that is zero-shot performance.
Why Modern LLMs Excel at Zero-Shot
Large language models trained on diverse, massive datasets develop broad capabilities that transfer to many tasks. This is why ChatGPT and similar models can answer questions, write code, translate languages, and analyse data without task-specific training—the breadth of pre-training enables zero-shot generalisation.
When to Use Zero-Shot vs. Few-Shot
Zero-shot is appropriate for simple, well-defined tasks where the model's pre-training likely covered similar examples. For specialised or unusual tasks—industry-specific terminology, unusual output formats, niche content styles—providing a few examples (few-shot) typically improves accuracy significantly.
Ready to close the loop?
See every term in action
Aergos tracks your AI and organic visibility across every channel, in one platform.
Not ready to talk? Audit your site free →