Cognition · Software component
Self-Reflection Critic
Software componentCognitionCognition & Memoryarc:SelfReflectionCritic
A reflection critic in which the same model that generated an output critiques and refines it.
Responsibility. Critiques the same model's own output.
Also known as: Single-agent self-reflection, Self-verification, Self-evaluation, Dual-pass reasoning verifier, Role-switching self-verifier, Constitutional self-critique
Variant of Reflection Critic abstract
When to choose. Choose when implementation simplicity matters and the generating model has adequate domain knowledge; beware it may reinforce its own misconceptions.
Relationships
is configured by structural
- Constitution abstract Ch9.6
- Critique Rubric Ch3.9 Ch9.5
invokes dependency
is invoked by dependency
writes dependency
escalates to dynamic
sends data to dynamic
is orchestrated by control
evaluates assurance
alternative to variability
Design guidance
- MUST be combined with premise validation; backward verification checks consistency, not premise truth.
- SHOULD separate generation from verification through distinct prompts or model calls to avoid the generator overlooking its own flaws.
- SHOULD use a structured step-by-step evaluation prompt rather than a generic 'check your work' request.
Quantitative guidance
As stated by the sources; verify before use.
- Self-evaluation roughly doubles computational cost (generate plus evaluate) (Ch3.9).
Classification
- Patterns
- Single-agent self-reflectionForward reasoning with backward verificationStep-by-step self-verificationSeparation of generation from verificationDual-pass reasoning architectureReflection and refinement (meta-reasoning over prior CoT)Iterative refinementPrinciple-guided self-critiqueConstitutional AI self-critique
- Quality attributes
- Maintainability (ISO/IEC 25010)
- Risks mitigated
- Anchoring to initial reasoning path
Sources
- Ch1.2: T. Nguyen, "Core Agent Patterns," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 1.2. ISBN: 9798244538229.
- Ch3.6: T. Nguyen, "Trace Analysis and Execution Debugging," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 3.6. ISBN: 9798244538229.
- Ch3.9: T. Nguyen, "Reasoning Quality," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 3.9. ISBN: 9798244538229.
- Ch5.1: T. Nguyen, "Chain-of-Thought (CoT) Fundamentals," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 5.1. ISBN: 9798244538229.
- Ch9.5: T. Nguyen, "Constitutional AI," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 9.5. ISBN: 9798244538229.
- Ch9.6: T. Nguyen, "Value Alignment Frameworks," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 9.6. ISBN: 9798244538229.