Prompt Evaluation Report

Use this template to evaluation summary for [prompt] or [LLM workflow].

Template Metadata

Evaluation Goal

Define the task, expected behavior, and release decision needed.

Notes

[Add context, assumptions, exceptions, evidence links, screenshots, calculations, or reviewer comments.]

Prompt Versions

Compare candidate prompts, model versions, parameters, and tool access.

Notes

[Add context, assumptions, exceptions, evidence links, screenshots, calculations, or reviewer comments.]

Test Set

Describe dataset size, source, sampling, sensitive cases, and holdout policy.

Notes

[Add context, assumptions, exceptions, evidence links, screenshots, calculations, or reviewer comments.]

Scoring Rubric

Define pass/fail criteria and weighted quality dimensions.

Notes

[Add context, assumptions, exceptions, evidence links, screenshots, calculations, or reviewer comments.]

Results

Summarize aggregate scores, segment performance, latency, and cost.

Notes

[Add context, assumptions, exceptions, evidence links, screenshots, calculations, or reviewer comments.]

Failures

Group notable failure modes with examples and severity.

Notes

[Add context, assumptions, exceptions, evidence links, screenshots, calculations, or reviewer comments.]

Recommendation

State the release decision, required changes, and monitoring plan. Keep examples concise and avoid exposing sensitive prompt secrets.

Notes

[Add context, assumptions, exceptions, evidence links, screenshots, calculations, or reviewer comments.]

Review and Signoff

Document review conclusions, approvals, unresolved items, and next review date.