ModelRefs / Agent Reflection — AI Glossary
Agent Reflection — AI Glossary
A pattern where an agent critiques its own outputs or reasoning to identify errors and self-correct before finalizing.
Overview
Reflection (also called self-critique) adds a dedicated review step: the agent generates an answer, then evaluates it against a rubric, and revises if needed. Substantially improves accuracy on tasks with verifiable correctness. Powers evaluator-optimizer agent workflows.
Reference details
| Topic | agents |
|---|---|
| Last reviewed | 2026-06-24 |
Related terms
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Agent Reflection — AI Glossary.
Frequently asked questions
What is Agent Reflection?
A pattern where an agent critiques its own outputs or reasoning to identify errors and self-correct before finalizing.
What concepts are related to Agent Reflection?
Closely related concepts include agent, agentic loop, eval.