Zero-shot prompting asks a language model to perform a task using only a description of what to do, without showing task-specific examples. It is the simplest prompting approach and often works well for everyday requests. section 3.2.2
A zero-shot request
Asking a model to summarize a document without providing a sample summary is zero-shot prompting. The model receives the task, but must infer the expected response from the instruction alone. section 3.2.2
Use zero-shot prompting as the starting point:
- Describe the task clearly and specify the desired response when the output is straightforward. section 3.2.2
- Switch to one-shot or few-shot prompting when the format, classification rules, style, or consistency are difficult to describe. section 3.2.2
- Treat the result as information to evaluate, because clear prompting does not remove hallucinations, overconfidence, bias, or other model limitations. section 3.5
Figure 3.4 Three common prompting approaches. In zero-shot prompting, the model receives only a task. In one-shot prompting, a single example is provided. In few-shot prompting, multiple examples are provided to establish a clearer pattern.

Compare zero-shot prompting with one-shot and few-shot prompting by the number of examples supplied to the model. section 3.2.2











