easyfundamentalspromptingZero-shot vs few-shot promptingWhat is the difference between zero-shot and few-shot prompting, and when would you choose one over the other?▾
easysystem-promptfundamentalsWhat is a system prompt?What is a system prompt and how does it differ from a user message? What should you put in it?▾
easysamplingparametersWhat does temperature control?Explain what the temperature parameter does when generating text with an LLM. What values would you use for different tasks?▾
easysamplingparametersTop-p and top-k samplingWhat are top-p (nucleus) sampling and top-k sampling? How do they differ from temperature?▾
easyrole-promptingfundamentalsRole promptingWhat is role prompting and why does it affect model outputs?▾
easyoutput-formatjsonRequesting structured outputHow do you reliably get an LLM to return JSON or another structured format?▾
easyhallucinationsafetyWhat is hallucination?What is hallucination in the context of LLMs, and why does it happen?▾
mediumchain-of-thoughtreasoningChain-of-thought (CoT) promptingExplain chain-of-thought prompting. How does it improve reasoning, and what are its variants?▾
mediumtree-of-thoughtreasoningTree-of-thought (ToT) promptingWhat is the tree-of-thought prompting framework, and how does it differ from chain-of-thought?▾
mediumself-consistencyreasoningdecodingSelf-consistency decodingWhat is self-consistency in prompting, and how does it improve answer reliability?▾
mediumsecurityprompt-injectionPrompt injection attacksWhat is a prompt injection attack, and what techniques can be used to defend against it?▾
mediumreactagentstool-useReAct (Reasoning + Acting) patternDescribe the ReAct prompting pattern. How does it combine reasoning and tool use?▾
mediumprompt-chainingdecompositionPrompt chaining and decompositionWhat is prompt chaining? When should you decompose a task into multiple prompts instead of one large prompt?▾
mediumragretrievalgroundingRetrieval-augmented promptingHow does retrieval-augmented generation (RAG) work at the prompting level? What are the key design decisions?▾
mediumhallucinationmitigationproductionHallucination mitigation strategiesWhat are the most effective strategies for reducing hallucinations in production LLM applications?▾
mediumcontext-windowtoken-limitsContext window managementWhat strategies do you use to manage prompts that approach or exceed the model's context window limit?▾
mediumcompressiontoken-limitsefficiencyPrompt compression techniquesWhat techniques can be used to compress prompts without losing important information?▾
hardalignmentrlhfconstitutional-aiConstitutional AI and RLHF alignmentWhat is Constitutional AI (CAI), how does it differ from RLHF, and how do alignment techniques affect prompt engineering?▾
hardsecurityadversarialrobustnessAdversarial robustness of promptsHow do you design prompts and systems that are robust to adversarial users trying to manipulate the model's behaviour?▾
hardmeta-promptingoptimisationMeta-promptingWhat is meta-prompting? How can you use a model to generate or improve its own prompts?▾
hardevaluationtestingmetricsEvaluating prompt qualityHow do you rigorously evaluate whether one prompt is better than another? What metrics and methodologies do you use?▾
hardgpt4claudegeminimodel-comparisonModel-specific prompting best practicesHow do prompting best practices differ between GPT-4, Claude, and Gemini? What model-specific considerations matter?▾
harda-b-testingproductionevaluationA/B testing prompts in productionHow do you set up a proper A/B test for prompts in a production LLM application?▾
hardleast-to-mostreasoningdecompositionLeast-to-most promptingWhat is least-to-most prompting and in what scenarios does it outperform standard chain-of-thought?▾
harddspyoptimisationprogrammaticProgrammatic prompt optimisation with DSPyWhat is DSPy and how does it change the approach to prompt engineering?▾
hardcostefficiencyproductiontoken-budgetToken budget and cost optimisationHow do you optimise prompt design to minimise token costs in a high-volume production system without sacrificing quality?▾