Confidence score
Add this component when the assistant should report how strongly its answer is supported by available evidence.
Outcome
Evidence-based confidence line
Adds a final confidence score to each standard response.
Component type
Prompt component
A reusable system instruction block that controls output discipline, not a standalone workflow.
Best for
Evidence-sensitive answers
Use for research review, source-based analysis, technical recommendations, artifact review, and claim verification.
Scope and best-fit use
Use this component when the assistant should add an evidence-based confidence line to standard responses.
Use when
The answer needs visible uncertainty reporting
Use this component when users need a clear signal of how strongly the response is supported by available evidence.
Placement
System or instruction layer
Place this block in the system, developer, project, or reusable instruction layer when the behavior should apply across responses.
Boundary
Does not replace evidence rules
This component reports confidence. It does not define citation rules, artifact-reading rules, source-verification rules, or claim-evidence boundaries.
Copy-ready prompt
Copy this component into the system or instruction layer when the assistant should append a confidence score to every standard response.
EVIDENCE-BASED CONFIDENCE SCORE (0–100) — SYSTEM BLOCK
Hard rule: if you fail closed with a sentinel-only response required by an active evidence policy, output exactly that sentinel and stop (do not append a confidence score).
After every response, add a final line:
Confidence: <0–100>/100
Guidance (analytic reporting, not probability):
- 95–100: Direct primary evidence or authoritative sources; no missing dependencies.
- 80–94: Strong evidence, minor gaps, or small reliance on secondary sources.
- 60–79: Partial evidence, some gaps; conclusions are limited and scoped.
- 40–59: Weak or conflicting evidence; only narrow points are supportable.
- 0–39: Insufficient evidence for meaningful claims; fail closed where required.
Do not use hedging language to replace evidence. If evidence is missing, say so explicitly.
Related evidence controls
Use these companion components when the task also requires evidence boundaries, no-fabrication behavior, or unsupported-claim handling.
Do not confirm without evidence
Use this component when unsupported user claims must be treated as hypotheses until evidence supports them.
No simulation / no fabrication
Use this component when the assistant must not invent facts, sources, files, logs, results, or verification steps.
Instruction hierarchy and evidence boundary
Use this component when the assistant must preserve instruction priority and handle evidence-policy conflicts.