← Master Index
Vol. 13 Module 13.1 Lecture

Reflection

Prompting Techniques

How This Lesson Fits the Module & Volume

Self-consistency improves answers by voting across samples. Reflection improves answers by critique: generate a draft, examine it against criteria, then revise. It is sequential and often cheaper than large-n sampling when a single thoughtful repair fixes the issue.

Reflection connects forward to prompt chaining (draft → critique → revise as stages) and to guardrails (policy checks as structured critique). It also complements role prompting when you cast separate generator and reviewer personas.

Learning Objectives

By the end of this lesson, students should be able to:

  • Define reflection (self-refine) prompting as draft → critique → revise.
  • Write critique rubrics that catch concrete failure modes.
  • Separate generator and reviewer prompts to reduce rubber-stamping.
  • Decide between reflection, self-consistency, and ToT for a workload.
  • Limit revision loops to avoid endless polish without quality gains.
  • Log critiques for debugging and evaluation.
Definition

Reflection (also called self-refine / self-critique) is a prompting pattern where the model produces an initial output, then generates explicit feedback against a checklist or rubric, and finally produces an improved revision that addresses that feedback.

The Three Passes

1. Draft

Solve the user task

2. Critique

Score vs rubric

3. Revise

Fix cited issues

Draft Prompt

Write a release note for engineers from the changelog bullets below. Audience: backend team. Length: 120–160 words. No marketing fluff. Changelog: {{BULLETS}}

Critique Prompt

You are a strict editor. Critique the draft against this rubric: 1) Factual coverage of every changelog bullet 2) Word count in 120–160 3) No marketing adjectives 4) Clear migration / action items if present in bullets Draft: {{DRAFT}} Return: - PASS or FAIL - bullet list of issues (empty if PASS) - missing changelog items (if any)

Revise Prompt

Revise the draft to resolve every issue below. Do not mention the critique. Output only the improved release note. Draft: {{DRAFT}} Issues: {{ISSUES}}

Reflection vs Related Patterns

PatternMechanismBest when
ReflectionCritique + revise one lineageWritable rubrics; fixable drafts
Self-consistencyVote across samplesDiscrete answers; diversity helps
ToTBranching searchHard planning / puzzles
CoT aloneOne reasoned passCheap multi-step baseline

Strengths and Tradeoffs

Strengths

  • Targets known failure modes via rubrics.
  • Often 2–3 calls beat large-n voting on prose.
  • Critiques create an audit trail.

Tradeoffs

  • Weak rubrics yield empty praise.
  • Models may “fix” issues without real change.
  • Extra latency versus single-shot.
Common Misconception

“Ask the model ‘are you sure?’ and that is reflection.” Vague self-checks produce vague confidence. Effective reflection uses a specific rubric, requires cited issues, and runs a dedicated revise step that must address those issues.

Knowledge Check

  1. Short Answer: What are the three stages of reflection? Answer: Draft, critique, revise.
  2. True/False: Reflection is the same as majority voting. Answer: False.
  3. Multiple Choice: A good critique prompt includes: (a) a concrete rubric, (b) only “be better,” (c) CUDA flags. Answer: (a).
  4. Short Answer: Why separate reviewer and writer roles? Answer: Reduces rubber-stamping; clearer critique focus.
  5. True/False: Endless revision loops always improve quality. Answer: False—gains diminish; cap iterations.
  6. Multiple Choice: Reflection often beats large-n voting for: (a) long prose with rubrics, (b) coin flips, (c) kernel convolution. Answer: (a).
  7. Short Answer: Name one failure mode of reflection. Answer: Empty praise / fake fixes / weak rubrics (any).
  8. True/False: Critiques can be logged for evaluation. Answer: True.
  9. Multiple Choice: Closest orchestration cousin: (a) prompt chaining, (b) max pooling, (c) byte-pair encoding. Answer: (a).
  10. Short Answer: What should the revise step output? Answer: An improved artifact that addresses cited issues (not a meta-discussion).

Key Takeaways

  • Reflection = draft, rubric-based critique, then revise.
  • Specific checklists beat vague “are you sure?” prompts.
  • Cap loops; verify that revisions actually change the output.
  • Next: Role Prompting.
Trainer’s Guide

Hands-on idea: Same draft with a vague critique vs a four-point rubric; compare revision quality.

Discussion prompt: When should critique be a separate model/role versus the same model in one multi-turn thread?

Recap: Reflection improves outputs through explicit critique and revision. Continue with Role Prompting.