A grelha de avaliação do Aristo
Versão rubric-2.0
Todas as pontuações do Aristo são calculadas a partir desta grelha. É pública para que possa discordar dela, e versionada para que o seu histórico continue comparável.
Como um prompt é pontuado
Cada critério é pontuado de 0 a 4 segundo as âncoras abaixo, ou marcado como não aplicável quando a tarefa o torna irrelevante. A pontuação global é a média ponderada, apresentada de 0 a 100. São recolhidas três amostras independentes do avaliador e prevalece a mediana; quando divergem, juntam-se mais duas amostras de um modelo mais potente e é apresentada uma nota de confiança.
Objetivo e critérios de sucesso
O resultado pretendido está indicado, e como saberemos que é bom?
Âncoras
- 0
No discernible outcome.
Exemplo: Tell me about marketing.
- 1
Topic or activity stated, outcome implicit.
Exemplo: Write about our new product.
- 2
Outcome stated, no success criteria.
Exemplo: Write a launch email for our product.
- 3
Outcome plus at least one testable criterion.
Exemplo: …that gets existing customers to book a demo; under 150 words.
- 4
Outcome, testable criteria, and priority among them when they conflict.
Exemplo: …Success = clear single call to action and a demo link above the fold; brevity matters more than covering every feature.
Contexto e dados de entrada
O modelo tem os factos, materiais, dados e a situação de que precisa, claramente delimitados?
Âncoras
- 0
No context at all; the model must guess the situation, the subject and the material.
Exemplo: Summarise this.
- 1
Situation named in passing; the material or facts the task needs are missing or only alluded to.
Exemplo: Summarise the meeting we had about the budget.
- 2
The needed material is present but undelimited or incomplete, so the model cannot tell input from instruction.
Exemplo: Summarise the meeting notes below for the finance team. Budget review 14 May: Ana raised…
- 3
Material is clearly delimited (tags, fences or headings) and the situation is explained; small gaps remain.
Exemplo: <notes>…</notes> These are notes from the 14 May budget review; summarise them for the finance team.
- 4
Complete and delimited material, background the model could not infer, and a statement of what it does not have.
Exemplo: <notes>…</notes> Background: this is the third review this quarter; last time we overspent on travel. Figures for June are not yet available, so do not estimate them.
Público e finalidade de uso
A quem se destina o resultado, e o que vão fazer com ele?
Âncoras
- 0
No reader or purpose can be inferred.
Exemplo: Explain photosynthesis.
- 1
Reader implied by the topic or the register, but never stated.
Exemplo: Explain photosynthesis simply.
- 2
Reader named; what they will do with the output is not.
Exemplo: Explain photosynthesis to a 12-year-old.
- 3
Reader and purpose of use both stated.
Exemplo: Explain photosynthesis to a 12-year-old preparing for a science quiz next week.
- 4
Reader, purpose and what they already know or need, so the level, emphasis and register follow from it.
Exemplo: …for a 12-year-old who has covered cells but not chemistry; she needs to answer short written questions, so favour precise terms over analogies.
Restrições e limites
Âmbito, o que é obrigatório e o que é proibido, extensão, fontes, segurança, o que fazer em caso de incerteza.
Âncoras
- 0
No boundaries of any kind.
Exemplo: Write a blog post about remote work.
- 1
One vague boundary, usually length or tone, nothing on scope or sources.
Exemplo: Write a short blog post about remote work.
- 2
Several concrete boundaries (length, scope, sources) but silent on what to do when uncertain.
Exemplo: …about 600 words, for managers, no product recommendations, cite at least two studies.
- 3
Concrete boundaries plus explicit must-nots and an instruction for the uncertain case.
Exemplo: …if you cannot find a study published after 2023, say so rather than inventing one.
- 4
Boundaries are complete, consistent and prioritised, stated as what to do rather than only what to avoid.
Exemplo: …Word count may stretch to 800 if needed to keep both studies; accuracy outranks length; mark any claim you are unsure of with “(verify)”.
Instruções da tarefa
Passos claros, ordenados e sem ambiguidade; uma tarefa ou tarefas explicitamente sequenciadas.
Âncoras
- 0
No actionable instruction; a topic or a fragment.
Exemplo: Customer churn Q3.
- 1
One ambiguous instruction, or several tasks tangled together with no order.
Exemplo: Look at the churn data and also fix the dashboard and tell me what to do.
- 2
A clear single task, or multiple tasks listed but their order or dependencies unstated.
Exemplo: Analyse the Q3 churn data and recommend three actions.
- 3
Clear, ordered steps or an explicit decomposition; each step has one verb and one object.
Exemplo: 1. Compute churn by plan. 2. Identify the plan with the biggest rise. 3. Propose three actions for that plan.
- 4
Ordered steps with the reasoning behind them, what to do if a step cannot be completed, and where to stop.
Exemplo: …Step 2 matters because we can only fund one intervention. If the data for a plan is missing, skip it and say so. Stop after the three actions; do not draft the rollout plan.
Contrato de saída
Formato, estrutura, esquema, extensão, língua e tom do resultado.
Âncoras
- 0
Nothing said about the form of the result.
Exemplo: Compare the two proposals.
- 1
A single loose hint such as “briefly” or “in a table”.
Exemplo: Compare the two proposals in a table.
- 2
Format and rough length stated; structure, language or tone left open.
Exemplo: Compare the two proposals in a table of at most ten rows, then a one-paragraph recommendation.
- 3
Format, structure, length and language or tone all stated and consistent with the goal.
Exemplo: …columns: criterion, proposal A, proposal B, winner. Write in Portuguese, neutral tone, no preamble.
- 4
A complete contract: schema or template, length, language, tone, and how edge cases appear in the output.
Exemplo: …Return JSON {"rows":[{"criterion","a","b","winner"}],"recommendation"}; if a criterion cannot be judged, set winner to "tie" and explain in recommendation.
Exemplos e referências
Exemplos resolvidos ou material de referência quando ajudam de facto; variados, sem sobreajuste.
Âncoras
- 0
Examples would clearly help (a specific format, style or classification) but none is given.
Exemplo: Classify each support ticket by urgency.
- 1
A reference is mentioned but not provided, or one example that is incomplete.
Exemplo: Classify each ticket by urgency, like we did last month.
- 2
One complete example, or several that all show the same case, so the model may over-fit it.
Exemplo: <example>Ticket: “Site down for all users” → urgent</example>
- 3
Two to five varied examples in a consistent, delimited format.
Exemplo: <examples><example>…→ urgent</example><example>…→ normal</example><example>…→ low</example></examples>
- 4
Varied examples that include an edge case or a counter-example, each with a short note on why.
Exemplo: …<example>“Invoice typo” → low, because nothing is blocked</example><example>“Cannot log in (one user)” → normal, not urgent: single user</example>
Papel e voz
Uma perspetiva ou especialização que muda o resultado; orientações de tom.
Âncoras
- 0
A perspective would change the output and none is given, or the role given contradicts the task.
Exemplo: Review this contract clause.
- 1
A generic label with no bearing on the output (“you are a helpful assistant”).
Exemplo: You are a helpful assistant. Review this contract clause.
- 2
A relevant role named, but its perspective is not used in the instructions.
Exemplo: You are a commercial lawyer. Review this contract clause.
- 3
A role whose expertise shapes the task, plus guidance on voice or stance.
Exemplo: As a commercial lawyer acting for the buyer, review this clause and flag risks to the buyer; be direct, no hedging.
- 4
Role, stance, voice and the standard to judge by, applied only where it changes the output.
Exemplo: …Judge against standard Canadian commercial practice; where the clause is market-standard say so in one line and move on.
Adequação ao modelo e higiene
Adequado à família de modelos alvo; sem contradições nem maiúsculas a gritar; material longo antes da pergunta; sem truques obsoletos.
Âncoras
- 0
Contradictory instructions, or shouting in capitals (“CRITICAL”, “MUST”), or a trick that harms current models (forced step-by-step, prefill-style openings).
Exemplo: CRITICAL: you MUST think step by step and NEVER make mistakes. Also be brief but cover everything.
- 1
No contradictions, but obsolete tricks or heavy emphasis remain, or long material comes after the question.
Exemplo: Think step by step. What are the risks? [20 pages of contract follow]
- 2
Plain, calm instructions; structure is weak (no delimiters) or the request ignores the target family’s conventions.
Exemplo: Here is the contract and here are my questions, mixed together in one paragraph.
- 3
Calm, consistent instructions; long material first and the question last; delimiters or tags separate parts; explains the why where it matters.
Exemplo: <contract>…</contract> Review the clause above. Keep the answer short because it goes into a client email.
- 4
Everything at 3, written for the chosen model family: structure and emphasis in the vendor’s recommended form, effort left to the model, nothing redundant.
Exemplo: Same prompt with <contract> and <questions> tags, a one-line role, no “think carefully” padding, and success criteria the model can check itself.
Tipos de tarefa
Os pesos mudam consoante o tipo de tarefa. Por exemplo, as transformações de código não precisam de público, e os prompts de imagem dão mais peso ao contrato de saída.
| Tipos de tarefa | Objetivo e critérios de sucesso | Contexto e dados de entrada | Público e finalidade de uso | Restrições e limites | Instruções da tarefa | Contrato de saída | Exemplos e referências | Papel e voz | Adequação ao modelo e higiene |
|---|---|---|---|---|---|---|---|---|---|
| Escrita | 3 | 3 | 3 | 2 | 2 | 2 | 1 | 1 | 1 |
| Análise | 3 | 3 | 2 | 2 | 2 | 2 | N/A | 1 | 1 |
| Programação | 3 | 3 | N/A | 3 | 2 | 2 | 1 | N/A | 1 |
| Investigação | 3 | 3 | 2 | 3 | 2 | 2 | N/A | 1 | 1 |
| Extração de dados | 3 | 3 | N/A | 2 | 2 | 3 | 1 | N/A | 1 |
| Criativo | 2 | 3 | 2 | 2 | 2 | 2 | 1 | 2 | 1 |
| Tutoria | 3 | 3 | 3 | 2 | 2 | 2 | 1 | 1 | 1 |
| Prompt de sistema de agente | 3 | 3 | 2 | 3 | 3 | 2 | 1 | 2 | 2 |
| Geração de imagens | 3 | 2 | N/A | 2 | 1 | 3 | N/A | N/A | 1 |
| Outro | 3 | 3 | 2 | 2 | 2 | 2 | 1 | 1 | 1 |
A grelha segue a evidência sobre a formação em prompts (requisitos antes da formulação) e as orientações atuais dos fornecedores para modelos de raciocínio; as fontes estão listadas na documentação do projeto.