> For the complete documentation index, see [llms.txt](https://docs.biobox.io/guide/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.biobox.io/guide/how-to/ask-biobox-with-claude/tools/get-evidence.md).

# get\_evidence — Get evidence for a scored entity

Get evidence for a scored entity

Get evidence for a scored entity

## What it does

Explain WHY an entity scores the way it does in a scoring report — its per-criterion evidence contributions: which evidence lines support the entity, and how much each one contributes.

**When to use:** the user asks "why is ERBB4 the #1 ALS target / which evidence supports it?" — the evidence behind a score; drilling into a single row returned by get\_report\_scores to see what drives that entity's score; you have a reportId (from list\_reports / where\_entity\_ranks) and an entity's displayName and want the breakdown.

**When NOT to use:** the user wants the model's FULL criteria list regardless of any one entity (use get\_reasoning\_model); the user wants the ranked list of top entities itself, not one entity's evidence (use get\_report\_scores); the user wants an entity's rank ACROSS models, not its evidence within one report (use where\_entity\_ranks).

**Parameters:**

* `reportId` — the scoring report's id (from list\_reports or where\_entity\_ranks) — which report to read evidence from
* `displayName` — the displayName of the scored entity, e.g. "ERBB4" — the entity whose evidence you want

**Returns:** An object { reportId, reportName, snapshot, object, totalScore, normScore, nonzeroCount, contributors }. When the entity isn't scored in the report `found: false` is returned (object null, contributors empty — not an error); an unknown reportId returns a 404 Not Found error (isError). object is the scored entity ({uuid, displayName, dbLabels}). totalScore is the entity's total weighted score; normScore is its 0–1 score within the report. contributors are the per-criterion evidence lines, SORTED by contribution (weightedScore) descending; each has uuid (the evidence line's id — match it against get\_reasoning\_model's evidenceLine\[].uuid to see that line's full definition: its query, scoring node, and weight), name, category, weight (the criterion's configured weight in the model), rawScore (the raw per-criterion value — shown as the 'scaled' column in the UI), scaledScore (the gateway's scaled, weight-independent per-criterion score), and weightedScore (the weight-factored contribution = scaledScore × weight — matches the UI's 'weighted'; quote this for 'how much'). weightedScore is a per-line magnitude for ranking the evidence, NOT an additive share of the total. nonzeroCount is how many criteria actually moved the score (the lines that explain it). snapshot + reportId are provenance — cite them so the breakdown is traceable.

## Parameters

| Parameter     | Type   | Required | Description                                                                                               |
| ------------- | ------ | -------- | --------------------------------------------------------------------------------------------------------- |
| `reportId`    | string | yes      | the scoring report's id (from list\_reports or where\_entity\_ranks) — which report to read evidence from |
| `displayName` | string | yes      | the displayName of the scored entity, e.g. "ERBB4" — the entity whose evidence you want                   |

## Response

```ts
{
  found?: boolean | undefined;
  reportId: string | null;
  reportName: string | null;
  snapshot: string | null;
  object: {
    uuid: string | null;
    displayName: string | null;
    dbLabels: string[];
  } | null;
  totalScore: number | null;
  normScore: number | null;
  nonzeroCount: number;
  contributors: Array<{
    uuid: string;
    name: string;
    category: string;
    weight: number | null;
    rawScore: number | null;
    scaledScore: number | null;
    weightedScore: number | null;
  }>;
}
```

## Notes

* Score fields verified live against the report's top-3 entities (ERBB4/SOD1/FIG4). Each maps to a source value from the gateway's per-criterion `evidenceCatScore` (joined to the criterion `weight` from `bipartiteGraph.evidence`): `rawScore` = `evidenceCatScore.rawScore` verbatim (the raw per-criterion value; the UI shows THIS as its "scaled" column — e.g. ERBB4·ClinVar 2.5929); `scaledScore` = `evidenceCatScore.score` verbatim (the gateway's scaled, weight-INDEPENDENT score — ERBB4·ClinVar 0.2591050493502868); `weightedScore` = `scaledScore × weight` (the weight-factored contribution; the UI's "weighted" — ERBB4·ClinVar 0.2591… × 1.5 = 0.38865757402543016).
* Per-line `weightedScore` does NOT sum to `totalScore` — the gateway aggregates the lines non-linearly (ERBB4: Σ weightedScore = 2.044 vs totalScore 1.327). Treat `weightedScore` as a per-line ranking magnitude.

## Example

```json
// get_evidence({"reportId":"CR-Ksa7pFW1SOnDyPQ0kd","displayName":"ERBB4"})
{
  "reportId": "CR-Ksa7pFW1SOnDyPQ0kd",
  "displayName": "ERBB4"
}
```

Response (arrays truncated):

```json
{
  "reportId": "CR-Ksa7pFW1SOnDyPQ0kd",
  "reportName": "ALS",
  "snapshot": "2026-03-02T18:09:07.223Z",
  "object": {
    "uuid": "ENSG00000178568",
    "displayName": "ERBB4",
    "dbLabels": [
      "Gene",
      "ProteinCoding"
    ]
  },
  "totalScore": 1.3274501206893674,
  "normScore": 0.36858877328971584,
  "nonzeroCount": 11,
  "contributors": [
    {
      "uuid": "5bwrtFVx1jN0LyXN3JvhF",
      "name": "HPO- Mendelian Genetic Association",
      "category": "Genetic Association",
      "weight": 1,
      "rawScore": 1,
      "scaledScore": 1,
      "weightedScore": 1
    },
    {
      "uuid": "bv1PS5NpbxR2N8vsLLvty",
      "name": "ClinVar Pathogenic Varient",
      "category": "Genetic Association",
      "weight": 1.5,
      "rawScore": 2.5928571428571425,
      "scaledScore": 0.2591050493502868,
      "weightedScore": 0.38865757402543016
    },
    {
      "uuid": "tDPrNZxhVKw3l5--W-hfT",
      "name": "Gene is differentially expressed in disease oligodendrocyte vs healthy oligodendrocyte cells",
      "category": "Single Cell Gene Expression",
      "weight": 0.8,
      "rawScore": 0.06582032562212475,
      "scaledScore": 0.16827108166724147,
      "weightedScore": 0.1346168653337932
    },
    "… 8 more"
  ]
}
```