Evaluations and datasets · Lesson 5 of 18 · 7 min

Run a baseline evaluation

A baseline evaluation freezes model, prompt, dataset and scoring configuration. A rerun is comparable only when the relevant conditions are recorded.

Open the lab Course map

Mastery 0/3Run itExplain itApply it

learner@clairos:~ $

↑ history · Tab complete · Ctrl+C cancel · Esc, Tab to leaveExit 0

Your objectives

Hints are always available. Wrong turns never cost XP.

1
Read the case evidence
Read /opt/ai-lab/cases/aiops-evals-datasets-005.json. Separate the recorded facts from your assumptions.
Hint & worked example
A baseline evaluation freezes model, prompt, dataset and scoring configuration. A rerun is comparable only when the relevant conditions are recorded.
The complete case JSON is read successfully.
```
cat /opt/ai-lab/cases/aiops-evals-datasets-005.json
```
Common mistake to avoid: Compare future scores without recording the scorer version.
2
Isolate the deciding signal
Select the scorer field from the case JSON. What does it establish—and what does it not establish?
Hint & worked example
Keep the full eval-22 version tuple before changing the candidate.
A successful JSON query isolates scorer.
```
jq .scorer /opt/ai-lab/cases/aiops-evals-datasets-005.json
```
Common mistake to avoid: Using a different case or an unrelated signal as the justification.
3
Record your operational decision
Choose the justified code: freeze-baseline-tuple, change-without-evidence, or ignore-the-signal. Save that one code followed by a newline in /home/learner/lab/aiops-evals-datasets-005-decision.txt.
Hint & worked example
Keep the full eval-22 version tuple before changing the candidate.
The decision artifact contains the correct code, not just a matching command in history.
```
printf '%s\n' 'freeze-baseline-tuple' > /home/learner/lab/aiops-evals-datasets-005-decision.txt
```
Common mistake to avoid: Compare future scores without recording the scorer version.

Understand the pattern

A baseline evaluation freezes model, prompt, dataset and scoring configuration. A rerun is comparable only when the relevant conditions are recorded.

Interpret the evidence

Keep the full eval-22 version tuple before changing the candidate.

A tempting mistake

Compare future scores without recording the scorer version.

All names, measurements, prices, versions, and thresholds in this exercise are fictional training data. The evidence supports this bounded scenario, not a claim about a live provider.

Worked example and expected result

$ cat /opt/ai-lab/cases/aiops-evals-datasets-005.json

{
  "fixture": "Fictional training data; not a live measurement.",
  "run": "eval-22",
  "model": "m3",
  "prompt": "p7",
  "dataset": "gold-v4",
  "scorer": "rubric-v2"
}

Exit status: 0. The command completed successfully in the training fixture.

$ jq .scorer /opt/ai-lab/cases/aiops-evals-datasets-005.json

"rubric-v2"

Exit status: 0. The command completed successfully in the training fixture.

$ printf '%s\n' 'freeze-baseline-tuple' > /home/learner/lab/aiops-evals-datasets-005-decision.txt

(no output; exit 0)

Exit status: 0. The command completed successfully in the training fixture.

Try a variation

Name the assumption that would have to change for “freeze-baseline-tuple” to stop being the justified action. Do not silently replace missing evidence with a guess.

Explain it

What does the evidence in “Run a baseline evaluation” justify?Keep the full eval-22 version tuple before changing the candidate.Compare future scores without recording the scorer version.Make the change now and gather the missing evidence afterward.

Apply it elsewhere

A colleague concludes: “Compare future scores without recording the scorer version.” What should you do?Accept it because the training command completed.Treat this fixture as a live measurement.Reject that conclusion; return to the case evidence and its stated limits.

Keep the momentum.

Complete the lab, then explain and apply the pattern.

Next: Read aggregate scores Try an incident drill

Anonymous lesson and mastery progress lasts for this session. Sign in to keep synced lesson XP between sessions; mastery stays device-local. No speed bonus, lost lives, or hint penalties.

Keep learning

Case methodology: evidence, limits, decision