> ## Documentation Index
> Fetch the complete documentation index at: https://trygradient.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Read an assessment scoring rubric

> Read the sub-criteria and 0-4 anchors inside a Gradient rubric. Gradient writes the rubric, and category weights stay fixed by version.

# Scoring Rubrics

Gradient's default rubric (v5) evaluates candidates across five categories (Correctness, Deliverable Quality, Reflection Quality, AI Fluency, Prioritized Skills) totaling 100 points. Gradient builds the rubric for each assessment. See [Scoring](/docs/concepts/scoring) for what each category measures.

## Default rubric

The overall score out of 100 is a weighted blend of the five categories. The **category weights are fixed by the rubric version**; Gradient sets the sub-criteria inside each category.

| Category | Weight | Scoring method |
| - | - | - |
| Correctness | 20% | Deterministic |
| Deliverable Quality | 20% | LLM Judge |
| Reflection Quality | 10% | LLM Judge |
| AI Fluency | 25% | Hybrid |
| Prioritized Skills | 25% | LLM Judge |

Each category contains sub-criteria. Every sub-criterion carries a point value, a short description, and a **0-4 anchored scale** (its `anchors`) that defines what each score from 0 to 4 means. The anchors are the scale the judge actually grades against, and they are what [calibration](#anchored-0-4-scales) reads and sharpens.

<Warning>
  **AI Fluency is locked.** Gradient manages it centrally. It cannot be edited or reweighted.
</Warning>

## Who sets the rubric

Gradient writes the rubric for each assessment when it builds the assessment. You cannot edit the
rubric. If you want a different rubric, email [hi@trygradient.ai](mailto:hi@trygradient.ai). You can still read the rubric on
the assessment page, and you can sharpen it through [calibration](/docs/concepts/scoring#calibration).

## Anchored 0-4 scales

Every sub-criterion should carry an `anchors` ladder: five descriptors, one for each score from 0 to 4, where 0 is absent or nothing and 4 is exceptional. Gradient writes them.

```json theme={null}
"anchors": [
  { "level": 0, "descriptor": "None of the required elements present." },
  { "level": 1, "descriptor": "One required element present." },
  { "level": 2, "descriptor": "About half the required elements present." },
  { "level": 3, "descriptor": "Most required elements present." },
  { "level": 4, "descriptor": "Every required element present." }
]
```

The anchors are not decoration. They matter in two ways:

* **They are the scoring scale.** The judge grades a sub-criterion against its anchor descriptors, not its one-line `description`. The description frames the criterion; the anchors define what each level actually means. A criterion with vague or missing anchors produces vague scores.
* **They are what calibration tunes.** When a reviewer grades a candidate on the 0-4 ladder and disagrees with the AI, [calibration](/docs/concepts/scoring#calibration) sharpens these anchor descriptors so future scoring matches your standard. A sub-criterion with no anchors cannot be calibrated, because there is no ladder to sharpen.

## Scoring methods

Sub-criteria are scored by one of these methods:

<CardGroup cols={2}>
  <Card title="Deterministic" icon="circle-check">
    Ground-truth checks against known-correct facts and required elements. Used for Correctness.
  </Card>

  <Card title="LLM Judge" icon="brain">
    An AI evaluator assesses quality against the rubric criteria. Best for subjective qualities like insight and reflection.
  </Card>

  <Card title="Event Analysis" icon="chart-line">
    Automated analysis of candidate behavior patterns. Best for measurable actions like whether a figure was verified before use.
  </Card>

  <Card title="Hybrid" icon="arrows-split-up-and-left">
    Combines approaches. AI Fluency is graded this way, centrally, by Gradient.
  </Card>
</CardGroup>

## Tips for reading a rubric

* **Read the anchors.** The anchors show what each score from 0 to 4 means for that sub-criterion.
* **Check the points.** Sub-criteria with more points have more effect on the total score.
* **Use calibration.** When you disagree with a score, grade the candidate yourself. Calibration then sharpens the anchors.
* **Do not try to edit AI Fluency.** It is locked.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.