Zero-Shot Learning
Asking a model to perform a task it has never seen demonstrated, relying purely on instructions.
Full Definition
Zero-shot prompting gives the model only a task description — no examples. The model must generalise from its pre-training knowledge and instruction tuning to perform the task correctly. Modern large language models are surprisingly capable zero-shot performers because instruction tuning teaches them to follow natural-language task descriptions across a wide variety of domains. Zero-shot prompting is the lowest-effort approach and works well for common NLP tasks (summarisation, translation, classification) but tends to underperform few-shot approaches on niche formats or unusual output schemas where the model has no prior reference point.
Examples
'Classify the sentiment of this tweet as positive, negative, or neutral: I can't believe how slow this app is today.'
'Translate the following sentence into Swahili: The meeting starts at 9 AM.'
Apply this in your prompts
PromptITIN automatically uses techniques like Zero-Shot Learning to build better prompts for you.
Related Terms
Few-Shot Learning
Providing a small number of input-output examples in the prompt to teach the mod…
View →Instruction Prompting
Directly telling the model what to do using clear imperative commands.…
View →Instruction-Tuned Model
A model fine-tuned on instruction-response pairs to follow natural-language dire…
View →