Arize Evaluator Skill

Concepts

What is an Evaluator?

An evaluator is an LLM-as-judge definition. It contains:

Evaluators are versioned — every prompt or model change creates a new immutable version. The most recent version is active.

What is a Task?

A task is how you run one or more evaluators against real data. Tasks are attached to a project (live traces/spans) or a dataset (experiment runs). A task contains:

Data Granularity

The --data-granularity flag controls what unit of data the evaluator scores. It defaults to span and only applies to project tasks (not dataset/experiment tasks — those evaluate experiment runs directly).

How trace and session aggregation works

For trace granularity, spans sharing the same context.trace_id are grouped together. Column values used by the evaluator template are comma-joined into a single string (each value truncated to 100K characters) before being passed to the judge model.

For session granularity, the same trace-level grouping happens first, then traces are ordered by start_time and grouped by attributes.session.id. Session-level values are capped at 100K characters total.

The {conversation} template variable

At session granularity, {conversation} is a special template variable that renders as a JSON array of {input, output} turns across all traces in the session, built from attributes.input.value / attributes.llm.input_messages (input side) and attributes.output.value / attributes.llm.output_messages (output side).

At span or trace granularity, {conversation} is treated as a regular template variable and resolved via column mappings like any other.

Multi-evaluator tasks

A task can contain evaluators at different granularities. At runtime the system uses the highest granularity (session > trace > span) for data fetching and automatically splits into one child run per evaluator. Per-evaluator query_filter in the task's evaluators JSON further narrows which spans are included (e.g., only tool-call spans within a session).

Basic CRUD

AI Integrations

AI integrations store the LLM provider credentials the evaluator uses. For full CRUD — listing, creating for all providers (OpenAI, Anthropic, Azure, Bedrock, Vertex, Gemini, NVIDIA NIM, custom), updating, and deleting — use the arize-ai-provider-integration skill.

Quick reference for the common case (OpenAI):

# Check for an existing integration first
ax ai-integrations list --space-id SPACE_ID

# Create if none exists
ax ai-integrations create \
  --name "My OpenAI Integration" \
  --provider openAI \
  --api-key $OPENAI_API_KEY

Copy the returned integration ID — it is required for ax evaluators create --ai-integration-id.

Evaluators

# List / Get
ax evaluators list --space-id SPACE_ID
ax evaluators get EVALUATOR_ID
ax evaluators list-versions EVALUATOR_ID
ax evaluators get-version VERSION_ID

# Create (creates the evaluator and its first version)
ax evaluators create \
  --name "Answer Correctness" \
  --space-id SPACE_ID \
  --description "Judges if the model answer is correct" \
  --template-name "correctness" \
  --commit-message "Initial version" \
  --ai-integration-id INT_ID \
  --model-name "gpt-4o" \
  --include-explanations \
  --use-function-calling \
  --classification-choices '{"correct": 1, "incorrect": 0}' \
  --template 'You are an evaluator. Given the user question and the model response, decide if the response correctly answers the question.

User question: {input}

Model response: {output}

Respond with exactly one of these labels: correct, incorrect'

# Create a new version (for prompt or model changes — versions are immutable)
ax evaluators create-version EVALUATOR_ID \
  --commit-message "Added context grounding" \
  --template-name "correctness" \
  --ai-integration-id INT_ID \
  --model-name "gpt-4o" \
  --include-explanations \
  --classification-choices '{"correct": 1, "incorrect": 0}' \
  --template 'Updated prompt...

{input} / {output} / {context}'

# Update metadata only (name, description — not prompt)
ax evaluators update EVALUATOR_ID \
  --name "New Name" \
  --description "Updated description"

# Delete (permanent — removes all versions)
ax evaluators delete EVALUATOR_ID

Key flags for create:

Tasks

# List / Get
ax tasks list --space-id SPACE_ID
ax tasks list --project-id PROJ_ID
ax tasks list --dataset-id DATASET_ID
ax tasks get TASK_ID

# Create (project — continuous)
ax tasks create \
  --name "Correctness Monitor" \
  --task-type template_evaluation \
  --project-id PROJ_ID \
  --evaluators '[{"evaluator_id": "EVAL_ID", "column_mappings": {"input": "attributes.input.value", "output": "attributes.output.value"}}]' \
  --is-continuous \
  --sampling-rate 0.1

# Create (project — one-time / backfill)
ax tasks create \
  --name "Correctness Backfill" \
  --task-type template_evaluation \
  --project-id PROJ_ID \
  --evaluators '[{"evaluator_id": "EVAL_ID", "column_mappings": {"input": "attributes.input.value", "output": "attributes.output.value"}}]' \
  --no-continuous

# Create (experiment / dataset)
ax tasks create \
  --name "Experiment Scoring" \
  --task-type template_evaluation \
  --dataset-id DATASET_ID \
  --experiment-ids "EXP_ID_1,EXP_ID_2" \
  --evaluators '[{"evaluator_id": "EVAL_ID", "column_mappings": {"output": "output"}}]' \
  --no-continuous

# Trigger a run (project task — use data window)
ax tasks trigger-run TASK_ID \
  --data-start-time "2026-03-20T00:00:00" \
  --data-end-time "2026-03-21T23:59:59" \
  --wait

# Trigger a run (experiment task — use experiment IDs)
ax tasks trigger-run TASK_ID \
  --experiment-ids "EXP_ID_1" \
  --wait

# Monitor
ax tasks list-runs TASK_ID
ax tasks get-run RUN_ID
ax tasks wait-for-run RUN_ID --timeout 300
ax tasks cancel-run RUN_ID --force

Time format for trigger-run: 2026-03-21T09:00:00 — no trailing Z.

Additional trigger-run flags:

Run status guide:

Workflow A: Create an evaluator for a project

Use this when the user says something like "create an evaluator for my Playground Traces project".

Step 1: Resolve the project name to an ID

ax spans export requires a project ID, not a name — passing a name causes a validation error. Always look up the ID first:

ax projects list --space-id SPACE_ID -o json

Find the entry whose "name" matches (case-insensitive). Copy its "id" (a base64 string).

Step 2: Understand what to evaluate

If the user specified the evaluator type (hallucination, correctness, relevance, etc.) → skip to Step 3.

If not, sample recent spans to base the evaluator on actual data:

ax spans export PROJECT_ID --space-id SPACE_ID -l 10 --days 30 --stdout

Inspect attributes.input, attributes.output, span kinds, and any existing annotations. Identify failure modes (e.g. hallucinated facts, off-topic answers, missing context) and propose 1–3 concrete evaluator ideas. Let the user pick.

Each suggestion must include: the evaluator name (bold), a one-sentence description of what it judges, and the binary label pair in parentheses. Format each like:

1. Name — Description of what is being judged. (label_a / label_b)

Example:

1. Response Correctness — Does the agent's response correctly address the user's financial query? (correct / incorrect)
2. Hallucination — Does the response fabricate facts not grounded in retrieved context? (factual / hallucinated)

Step 3: Confirm or create an AI integration

ax ai-integrations list --space-id SPACE_ID -o json

If a suitable integration exists, note its ID. If not, create one using the arize-ai-provider-integration skill. Ask the user which provider/model they want for the judge.

Step 4: Create the evaluator

Use the template design best practices below. Keep the evaluator name and variables generic — the task (Step 6) handles project-specific wiring via column_mappings.

ax evaluators create \
  --name "Hallucination" \
  --space-id SPACE_ID \
  --template-name "hallucination" \
  --commit-message "Initial version" \
  --ai-integration-id INT_ID \
  --model-name "gpt-4o" \
  --include-explanations \
  --use-function-calling \
  --classification-choices '{"factual": 1, "hallucinated": 0}' \
  --template 'You are an evaluator. Given the user question and the model response, decide if the response is factual or contains unsupported claims.

User question: {input}

Model response: {output}

Respond with exactly one of these labels: hallucinated, factual'

Step 5: Ask — backfill, continuous, or both?

Before creating the task, ask:

"Would you like to: (a) Run a backfill on historical spans (one-time)? (b) Set up continuous evaluation on new spans going forward? (c) Both — backfill now and keep scoring new spans automatically?"

Step 6: Determine column mappings from real span data

Do not guess paths. Pull a sample and inspect what fields are actually present:

ax spans export PROJECT_ID --space-id SPACE_ID -l 5 --days 7 --stdout

For each template variable ({input}, {output}, {context}), find the matching JSON path. Common starting points — always verify on your actual data before using:

Validate span kind alignment: If the evaluator prompt assumes LLM final text but the task targets CHAIN spans (or vice versa), runs can cancel or score the wrong text. Make sure the query_filter on the task matches the span kind you mapped.

Full example --evaluators JSON:

[
  {
    "evaluator_id": "EVAL_ID",
    "query_filter": "span_kind = 'LLM'",
    "column_mappings": {
      "input": "attributes.input.value",
      "output": "attributes.llm.output_messages.0.message.content",
      "context": "attributes.retrieval.documents.contents"
    }
  }
]

Include a mapping for every variable the template references. Omitting one causes runs to produce no valid scores.

Step 7: Create the task

Backfill only (a):

ax tasks create \
  --name "Hallucination Backfill" \
  --task-type template_evaluation \
  --project-id PROJECT_ID \
  --evaluators '[{"evaluator_id": "EVAL_ID", "column_mappings": {"input": "attributes.input.value", "output": "attributes.output.value"}}]' \
  --no-continuous

Continuous only (b):

ax tasks create \
  --name "Hallucination Monitor" \
  --task-type template_evaluation \
  --project-id PROJECT_ID \
  --evaluators '[{"evaluator_id": "EVAL_ID", "column_mappings": {"input": "attributes.input.value", "output": "attributes.output.value"}}]' \
  --is-continuous \
  --sampling-rate 0.1

Both (c): Use --is-continuous on create, then also trigger a backfill run in Step 8.

Step 8: Trigger a backfill run (if requested)

First find what time range has data:

ax spans export PROJECT_ID --space-id SPACE_ID -l 100 --days 1 --stdout   # try last 24h first
ax spans export PROJECT_ID --space-id SPACE_ID -l 100 --days 7 --stdout   # widen if empty

Use the start_time / end_time fields from real spans to set the window. Use the most recent data for your first test run.

ax tasks trigger-run TASK_ID \
  --data-start-time "2026-03-20T00:00:00" \
  --data-end-time "2026-03-21T23:59:59" \
  --wait

Workflow B: Create an evaluator for an experiment

Use this when the user says something like "create an evaluator for my experiment" or "evaluate my dataset runs".

If the user says "dataset" but doesn't have an experiment: A task must target an experiment (not a bare dataset). Ask:

"Evaluation tasks run against experiment runs, not datasets directly. Would you like help creating an experiment on that dataset first?"

If yes, use the arize-experiment skill to create one, then return here.

Step 1: Resolve dataset and experiment

ax datasets list --space-id SPACE_ID -o json
ax experiments list --dataset-id DATASET_ID -o json

Note the dataset ID and the experiment ID(s) to score.

Step 2: Understand what to evaluate

If the user specified the evaluator type → skip to Step 3.

If not, inspect a recent experiment run to base the evaluator on actual data:

ax experiments export EXPERIMENT_ID --stdout | python3 -c "import sys,json; runs=json.load(sys.stdin); print(json.dumps(runs[0], indent=2))"

Look at the output, input, evaluations, and metadata fields. Identify gaps (metrics the user cares about but doesn't have yet) and propose 1–3 evaluator ideas. Each suggestion must include: the evaluator name (bold), a one-sentence description, and the binary label pair in parentheses — same format as Workflow A, Step 2.

Step 3: Confirm or create an AI integration

Same as Workflow A, Step 3.

Step 4: Create the evaluator

Same as Workflow A, Step 4. Keep variables generic.

Step 5: Determine column mappings from real run data

Run data shape differs from span data. Inspect:

ax experiments export EXPERIMENT_ID --stdout | python3 -c "import sys,json; runs=json.load(sys.stdin); print(json.dumps(runs[0], indent=2))"

Common mapping for experiment runs:

• output → "output" (top-level field on each run)
• input → check if it's on the run or embedded in the linked dataset examples

If input is not on the run JSON, export dataset examples to find the path:

ax datasets export DATASET_ID --stdout | python3 -c "import sys,json; ex=json.load(sys.stdin); print(json.dumps(ex[0], indent=2))"

Step 6: Create the task

ax tasks create \
  --name "Experiment Correctness" \
  --task-type template_evaluation \
  --dataset-id DATASET_ID \
  --experiment-ids "EXP_ID" \
  --evaluators '[{"evaluator_id": "EVAL_ID", "column_mappings": {"output": "output"}}]' \
  --no-continuous

Step 7: Trigger and monitor

ax tasks trigger-run TASK_ID \
  --experiment-ids "EXP_ID" \
  --wait

ax tasks list-runs TASK_ID
ax tasks get-run RUN_ID

Best Practices for Template Design

1. Use generic, portable variable names

Use {input}, {output}, and {context} — not names tied to a specific project or span attribute (e.g. do not use {attributes_input_value}). The evaluator itself stays abstract; the task's column_mappings is where you wire it to the actual fields in a specific project or experiment. This lets the same evaluator run across multiple projects and experiments without modification.

2. Default to binary labels

Use exactly two clear string labels (e.g. hallucinated / factual, correct / incorrect, pass / fail). Binary labels are:

• Easiest for the judge model to produce consistently
• Most common in the industry
• Simplest to interpret in dashboards

If the user insists on more than two choices, that's fine — but recommend binary first and explain the tradeoff (more labels → more ambiguity → lower inter-rater reliability).

3. Be explicit about what the model must return

The template must tell the judge model to respond with only the label string — nothing else. The label strings in the prompt must exactly match the labels in --classification-choices (same spelling, same casing).

Good:

Respond with exactly one of these labels: hallucinated, factual

Bad (too open-ended):

Is this hallucinated? Answer yes or no.

4. Keep temperature low

Pass --invocation-params '{"temperature": 0}' for reproducible scoring. Higher temperatures introduce noise into evaluation results.

5. Use --include-explanations for debugging

During initial setup, always include explanations so you can verify the judge is reasoning correctly before trusting the labels at scale.

6. Pass the template in single quotes in bash

Single quotes prevent the shell from interpolating {variable} placeholders. Double quotes will cause issues:

# Correct
--template 'Judge this: {input} → {output}'

# Wrong — shell may interpret { } or fail
--template "Judge this: {input} → {output}"

7. Always set --classification-choices to match your template labels

The labels in --classification-choices must exactly match the labels referenced in --template (same spelling, same casing). Omitting --classification-choices causes task runs to fail with "missing rails and classification choices."

Troubleshooting

Related Skills

• arize-ai-provider-integration: Full CRUD for LLM provider integrations (create, update, delete credentials)
• arize-trace: Export spans to discover column paths and time ranges
• arize-experiment: Create experiments and export runs for experiment column mappings
• arize-dataset: Export dataset examples to find input fields when runs omit them
• arize-link: Deep links to evaluators and tasks in the Arize UI

Save Credentials for Future Use

See references/ax-profiles.md § Save Credentials for Future Use.