ARC2 PEEC logo
ARC2 PEEC™Prompt Engineering Evaluation Center · V2.0
ARC2 PEEC™ · PUBLIC EVALUATION

Employer Evaluation Center

Evaluate the prompt engineering workflow, test a transparent scoring rubric, review role requirements, and export your assessment. This public demonstration uses illustrative examples and never reveals private user records.

Role requirements

Design, test, refine and document enterprise prompts for GPT, Claude and comparable LLMs. Demonstrate accuracy, safety, relevance, performance, RAG, agent workflows, and governance.

  • Zero-shot, few-shot and role-based prompting
  • CoT/ToT/ReAct concepts and verifiable outcomes
  • Python, evaluation frameworks and agent orchestration
  • RAG, vector search, citation quality and grounding
  • Regression testing, safety, reproducibility and reporting

Evaluation workflow

  1. Submit prompt and recorded model response.
  2. Run five deterministic prompt-structure checks.
  3. Authenticator reviews the evidence and grades 0–100.
  4. Compare structural score and human critique.
  5. Archive authorized records and review aggregate metrics.

Live model execution and automated correctness judging are not included in this foundation.

Try the scoring methodology

Public example only. This score measures presence of prompt-design elements, not factual accuracy, safety or model performance.

Employer critique & scorecard

Use 0–5 ratings, or select Not assessed. Ratings are your independent assessment and are not sent to ARC2 automatically.

Implementation status

CapabilityStatus
Accounts, role access, archive and manual response captureIncluded in V2.0 foundation
Five structural prompt checks and human reviewIncluded; heuristic only
Public demonstration, employer rubric and exportIncluded in V2.2
GPT/Claude API execution, automated factual judgingNot implemented
Vector retrieval, agent trace grading and certificationNot implemented
External penetration testing and production certificationNot performed