Recipe catalog / answer-disclosures
Label the disclosures in an answer
Which of these does draft include: stated uncertainty, stated limitations, cited sources, stated assumptions?
You need to audit or gate model answers on transparency habits in a single call, for evaluation sets or response policies.
Explore this recipe interactively ยท Source and implementation guide
Use answer-disclosures in TypeScript
Install with npm install jev-recipes. Requires Node.js 22.9 or newer and ES modules. Set TYPESAFE_API_KEY in your server environment for live calls, which send input to TypeSafe and use API quota. See the installation guide.
import { answerDisclosures } from 'jev-recipes/answer-disclosures';
const result = await answerDisclosures({
"draft": "Based on the 2024 pricing page, the Team plan includes SSO. I am not certain this still applies to accounts created before March, and I have not checked the enterprise tier. I am assuming you are on a monthly plan.",
"minConfidence": 0.8
});
console.log(result);Input contract
| Field | Type | Needed |
|---|---|---|
| draft | string | Required |
| minConfidence | number | Optional |
Full input and result schemas
{
"input": {
"$schema": "https://json-schema.org/draft/2020-12/schema",
"type": "object",
"properties": {
"draft": {
"type": "string"
},
"minConfidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"draft"
]
},
"result": {
"$schema": "https://json-schema.org/draft/2020-12/schema",
"type": "object",
"properties": {
"model": {
"type": "string"
},
"usage": {
"type": "object",
"properties": {
"input_tokens": {
"type": "integer",
"minimum": 0,
"maximum": 9007199254740991
},
"output_tokens": {
"type": "integer",
"minimum": 0,
"maximum": 9007199254740991
}
},
"required": [
"input_tokens",
"output_tokens"
],
"additionalProperties": false
},
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"detected": {
"type": "array",
"items": {
"type": "string",
"enum": [
"statesUncertainty",
"statesLimitations",
"citesSources",
"statesAssumptions"
]
}
},
"labels": {
"type": "object",
"properties": {
"statesUncertainty": {
"type": "object",
"properties": {
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"verdict": {
"type": "string",
"enum": [
"present",
"absent"
]
},
"probability": {
"type": "number",
"minimum": 0,
"maximum": 1
},
"confidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"status",
"verdict",
"probability",
"confidence"
],
"additionalProperties": false
},
"statesLimitations": {
"type": "object",
"properties": {
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"verdict": {
"type": "string",
"enum": [
"present",
"absent"
]
},
"probability": {
"type": "number",
"minimum": 0,
"maximum": 1
},
"confidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"status",
"verdict",
"probability",
"confidence"
],
"additionalProperties": false
},
"citesSources": {
"type": "object",
"properties": {
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"verdict": {
"type": "string",
"enum": [
"present",
"absent"
]
},
"probability": {
"type": "number",
"minimum": 0,
"maximum": 1
},
"confidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"status",
"verdict",
"probability",
"confidence"
],
"additionalProperties": false
},
"statesAssumptions": {
"type": "object",
"properties": {
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"verdict": {
"type": "string",
"enum": [
"present",
"absent"
]
},
"probability": {
"type": "number",
"minimum": 0,
"maximum": 1
},
"confidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"status",
"verdict",
"probability",
"confidence"
],
"additionalProperties": false
}
},
"required": [
"statesUncertainty",
"statesLimitations",
"citesSources",
"statesAssumptions"
],
"additionalProperties": false
}
},
"required": [
"model",
"usage",
"status",
"detected",
"labels"
],
"additionalProperties": false
}
}Saved example result
This hand-authored response demonstrates the contract. It is not a model accuracy measurement. Run it without an API key: npx jev-recipes demo answer-disclosures.
{
"model": "demo-fixture",
"usage": {
"input_tokens": 0,
"output_tokens": 0
},
"status": "ready",
"detected": [
"statesUncertainty",
"statesLimitations",
"citesSources",
"statesAssumptions"
],
"labels": {
"statesUncertainty": {
"status": "ready",
"verdict": "present",
"probability": 0.95,
"confidence": 0.95
},
"statesLimitations": {
"status": "ready",
"verdict": "present",
"probability": 0.9,
"confidence": 0.9
},
"citesSources": {
"status": "ready",
"verdict": "present",
"probability": 0.88,
"confidence": 0.88
},
"statesAssumptions": {
"status": "ready",
"verdict": "present",
"probability": 0.94,
"confidence": 0.94
}
}
}
Evaluation evidence
No verified live accuracy measurement is available. Evaluate representative cases before using this decision in your workflow.
Use the evaluation guide to measure this decision on your own labeled cases.
Limitations
- Detects that a disclosure is present, not that it is accurate or sufficient.
- A cited source is detected by attribution wording. Verify the source exists in code.
Related recipes
- uncertainty-expression: Use uncertainty-expression to grade how certain a single claim sounds.
- citation-needed: Use citation-needed to decide whether a statement requires evidence under your rules.