Chatgpt Prompting Essentials

A prompt is the natural-language input—question, instruction, or context—a user provides to a language model to elicit a response. It frames the task and guides the model's behavior and output format.

The role pattern assigns the model a persona or function (e.g., 'You are a senior copyeditor'). It conditions tone, vocabulary, and reasoning style, often improving relevance for domain-specific tasks.

Few-shot prompting supplies a small number of input–output examples within the prompt before the new query, demonstrating the desired pattern, format, or reasoning so the model can generalize from the demonstrations.

Zero-shot prompting asks the model to perform a task using only the instructions in the prompt, with no examples. It relies on the model's pretrained knowledge to interpret and complete the request.

Chain-of-thought prompting encourages the model to produce intermediate reasoning steps before the final answer, improving accuracy on multi-step problems in arithmetic, logic, and commonsense tasks.

Self-consistency samples multiple independent chain-of-thought rationales for the same prompt and selects the most frequent final answer, reducing variance and improving accuracy on reasoning benchmarks.

A hallucination is a model-generated statement that is fluent and plausible but factually incorrect, fabricated, or unsupported by the source or training data, often presented with high confidence.

Grounding responses with retrieved documents (RAG), explicitly instructing the model to say 'I don't know' when uncertain, and requiring citations or quoted evidence for factual claims.

RAG combines a retriever that fetches relevant external documents with a generator (LLM) that conditions its answer on those documents, improving factual accuracy and allowing up-to-date knowledge.

The system message sets high-level behavior, persona, constraints, and formatting rules for the entire conversation, taking priority over later user instructions within the same thread.

Temperature scales the logits before sampling; higher values increase randomness. Top-p (nucleus) sampling restricts the token pool to the smallest set whose cumulative probability exceeds p.

The 'as if' framing asks the model to respond 'as if' it were a specific expert or in a specific scenario, leveraging role conditioning to elicit more accurate, context-appropriate answers.