Prompt Engineering Team Challenge

What we verified before publishing

Fact-check status: passed

The following claims were verified through Perplexity Sonar before the questions were finalised. The host can reference any of them during play:

• Prompt engineering is the process of structuring natural language inputs, known as prompts, to produce specified outputs from a generative artificial intelligence model[4].
• A prompt is the input provided to an AI model to elicit a specific response, such as an instruction that determines the content the model generates[1][4].
• Zero-shot prompting is the most direct method of prompt engineering where an AI is given a direct instruction or question without additional examples or information[2].
• Few-shot prompting involves supplying the generative AI with specific examples to help guide and improve the accuracy of its output[2].
• Prompt chaining is a technique that splits a complex task into smaller subtasks and uses the AI's outputs to accomplish the overarching goal[2].
• The best practice for prompt engineering is to express queries as clearly as possible using specific, plain language and context[2].
• Iterative refinement is a key workstep where engineers test prompts, evaluate outputs, and adjust the prompt until the desired quality is achieved[3][8].
• Clarity is key in prompt engineering, requiring prompts to be unambiguous and avoiding unnecessary jargon[8].
• Using constraints, such as limiting the response length to three sentences, helps guide the model toward the desired output format[8].
• Role-playing prompts, where the model assumes a specific role, yield more tailored and relevant responses for the user[8].
• Good prompt engineering requires communicating instructions with clear context, scope, and expected response criteria[6].
• Keywords strongly influence a model's understanding, so using the right terms like "mathematics" ensures correct contextual definitions[7].