Definition
Zero-shot prompting presents a task to a language model using nothing but an instruction — no worked examples are provided. The model must rely entirely on the general knowledge and patterns absorbed during pre-training to interpret the request and produce an appropriate answer. The term “zero-shot” refers to the zero demonstrations supplied in the prompt.
How it works
Asking a model “What is the capital of France?” or “Classify the sentiment of this review as positive or negative” are zero-shot prompts. The model already understands concepts like “capital” and “sentiment” from training, so it can respond without being shown labelled examples first. Modern instruction-tuned models are deliberately trained to follow such direct commands, which makes zero-shot prompting remarkably effective for everyday tasks.
Trade-offs
Zero-shot prompting is the simplest and cheapest approach: prompts are short, and you spend no tokens on examples. Its weakness appears when a task is unusual, requires a specific output format, or sits at the edge of the model's knowledge. In those cases, adding demonstrations — moving to one-shot prompting or few-shot prompting — often improves accuracy.
A practical workflow is to start zero-shot and only escalate to examples when results fall short. For sovereign Bitcoiners running models locally, beginning with the leanest prompt conserves both context-window space and compute. Zero-shot prompting is a cornerstone technique within wider prompt engineering.
In Simple Terms
Zero-shot prompting presents a task to a language model using nothing but an instruction — no worked examples are provided. The model must rely entirely…