48.1 Key terms: tokens, context, temperature, tools, RAG

On this page

Key Terms

Term	Definition
Token	A chunk of text (~4 characters). Models count input/output in tokens.
Context Window	Maximum tokens the model can process at once (e.g., 128K, 1M).
Temperature	Randomness in output. 0 = deterministic, 1 = creative.
Top-K	Only consider the top K most likely tokens.
Top-P	Only consider tokens with cumulative probability P.
System Instruction	Prompt that defines the model's behavior for all turns.
Few-Shot	Including examples in the prompt to teach the model.
Zero-Shot	Asking the model without examples.
Chain-of-Thought	Asking the model to think step-by-step.
RAG	Retrieval-Augmented Generation. Inject retrieved docs into prompt.
Embedding	Vector representation of text for similarity search.
Function Calling	Model can request to call external functions/tools.
Grounding	Connecting model output to real data sources.
Hallucination	Model confidently stating false information.
Prompt Injection	Malicious input that hijacks model behavior.