Model Ops and Safety

These concepts help teams monitor model behavior, control unsafe outputs, and keep humans involved in sensitive decisions.

How to recognize this theme

AI operations vocabulary for context, oversight, and behavior changes.

In a daily board, this category groups terms by their shared role. Look for four cards that describe the same mechanism, risk area, or workflow rather than four words that merely sound similar.

Educational context

These entries are vocabulary notes for learning. They are not project endorsements, token recommendations, exchange rankings, or trading signals.

Context Window

A context window is the amount of input and generated text a model can consider at one time when producing a response.

Guardrail

A guardrail is a policy, filter, prompt rule, or system check that helps keep model outputs within intended safety and quality boundaries.

Model Drift

Model drift is a change in model performance or behavior as input patterns, data quality, user needs, or surrounding systems evolve.

Human-in-the-Loop

Human-in-the-loop describes a workflow where people review, approve, correct, or override automated model decisions in important cases.