Engineering instrument

AI Hallucination Prompt Review Tool

Explore the sinc-LLM six-field prompt format: role, context, data, constraints, output format, and task. Review the structured prompt and verify its output.

Book a 30-minute call Compare services

The services, methods, and examples described here do not establish an expected performance result for your workload.

Review the prompt’s role, context, data, constraints, format, and task with the sinc-LLM six-field format. Use output checks and source verification when assessing hallucination.

Hallucination and Prompt Specification

This guide examines missing prompt fields as one part of hallucination review. Prompt structure alone does not establish factual correctness.

The author uses a signal-sampling analogy to discuss prompt specification. The six fields below organize task details; the analogy does not establish reliable reconstruction of intent. Related: sinc-LLM.

x(t) = Σ x(nT) · sinc((t - nT) / T)

Six Fields for Prompt Review

The sinc-LLM framework organizes prompt review into specification bands:

PERSONA — State the role, expertise, and communication style to request.
CONTEXT — State the environment and situation rather than leaving them implicit.
DATA — Specify the supplied inputs and their sources.
CONSTRAINTS — State rules, limits, and exclusions, including what the model may infer.
FORMAT — Specify the required output structure and schema.
TASK — State the objective explicitly.

Prompt Structure: An Illustrative Comparison

Example: Short Prompt

"Tell me about the environmental impact of lithium mining"

This example leaves the role, evidence boundary, and requested format unstated. Review the answer’s sources and factual claims.

Example: Six-Field Prompt

6-band decomposition: environmental scientist persona, lithium triangle context (Argentina/Bolivia/Chile), 2020-2026 data range, peer-reviewed sources only constraint, technical report format, specific analysis task with water usage focus.

A structured prompt makes its instructions explicit. It still requires checks against the task, supplied sources, and model output.

How to Review Prompt Fields

Paste your raw prompt into sinc-LLM
The tool automatically decomposes it into 6 bands
3. Review missing or underspecified fields against the task.
Fill in the missing bands with your specific requirements
Use the complete sinc JSON as your prompt

Review the proposed fields before use and check the model’s output against the task and supplied sources.

{
  "formula": "x(t) = \u03a3 x(nT) \u00b7 sinc((t - nT) / T)",
  "T": "specification-axis",
  "fragments": [
    {"n": 0, "t": "PERSONA", "x": "Expert data scientist with 10 years ML experience"},
    {"n": 1, "t": "CONTEXT", "x": "Building a recommendation engine for an e-commerce platform"},
    {"n": 2, "t": "DATA", "x": "Dataset: 2M user interactions, 50K products, sparse matrix"},
    {"n": 3, "t": "CONSTRAINTS", "x": "Must use collaborative filtering. Latency under 100ms. No PII in logs. Python 3.11+. Must handle cold-start users with content-based fallback"},
    {"n": 4, "t": "FORMAT", "x": "Python module with type hints, docstrings, and pytest tests"},
    {"n": 5, "t": "TASK", "x": "Implement the recommendation engine with train/predict/evaluate methods"}
  ]
}

Using the Format Across Models

The six-field format can be used when writing prompts for ChatGPT, Claude, Gemini, or a local model. Output behavior still depends on the model, task, sources, and settings.

Keep factual verification in the workflow. Use explicit prompt fields as one part of preparing and checking the task. Related: Try sinc-LLM now.

Open the Prompt Tool →. Related: Fix Hallucinations Free →.

Additional context

sinc-LLM About Blog Services