Max

Finds the maximum value of a score (with the given name) across all samples. Use this to surface best-case performance - for example, the highest judge score across all responses. Pair with Min and Mean for a complete picture of score distribution.

Output

A single metric named {field}_max by default, or the value of name if provided.

Examples

Example: RAG QA. Finds the highest judge score across all samples to identify the best-performing response in the run.

Max Metric

# ...
config_spec:
  - type: "dataset"
    key: "dataset_key"
    display_name: "Dataset"
    description: "The dataset to evaluate against."
  - type: "model"
    key: "judge_model_key"
    display_name: "Judge Model"
    description: "Model used to score the evaluation dimension."
  - type: "categorical"
    key: "evaluation_dimension"
    display_name: "Evaluation Dimension"
    description: "The aspect of the model response to evaluate."
    allowed_values: ["groundedness", "relevance"]
definition:
  # ...
  scorers:
    - type: "model_as_a_judge_scorer"
      model_key: "<< config.judge_model_key >>"
      system_prompt: >
        You are an evaluator assessing the << config.evaluation_dimension >> of a
        model response to a question.
        On a scale from 0 to 100, assign 100 if the response is fully
        << config.evaluation_dimension >> and 0 if it is not at all.
        Return only the numeric score.
      user_prompt: >
        <context>{{ sample.context }}</context>
        <response>{{ trace.get_last_assistant_text() }}</response>
      score_min: 0
      score_max: 100
      metrics:
        - type: "mean"
          field: "score"
          name: "Mean Score"
        - type: "min"
          field: "score"
          name: "Min Score"
        - type: "max"
          field: "score"
          name: "Max Score"

Configuration

Properties

---

type Literal “max” required

The type of the metric.

---

field string, TemplateValue required

The field over which to compute the max.

---

name string, TemplateValue

The name given to the metric value. If not specified, it is {field}_max

Default: None

---

key string

Unique identifier assigned to the entity in AI Platform.

Default: None