> For the complete documentation index, see [llms.txt](https://docs.neurolabai.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.neurolabai.io/start-using-neurolab/reports-and-raw-payload.md).

# Reports and Raw Payload

Neurolab presents one structured result through several coordinated views. The formatted report is optimized for reading; the raw payload is the authoritative representation of the fields actually returned.

The report can combine a stable predictive-model release with a newer facts-and-evidence release. Preserve both identities: model output and the supporting evidence layer can evolve on different schedules.

## Two views of the same artifact

The report cards summarize common sections such as mechanism of action, cognition, reported usage, side effects, ATC classification, brain outputs, and reliability. The raw JSON preserves exact keys, arrays, nulls, route metadata, support bundles, and version fields.

![Donepezil Raw Payload workspace with search copy and export controls](https://3908477117-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fgze1K0yofnhFNmM9nEyt%2Fuploads%2Fgit-blob-e36b042bc0a0206e3efe18df3b1b2ea807df2d20%2F03-raw-payload.jpg?alt=media)

If a card label and the payload appear to differ, inspect the key that backs the card. For example, a displayed “Reported Usage” section can be backed by `reported_usage_top_k` or a compatible fallback field. Do not infer a field that is not present.

## Ranked lists

Many predicted sections use a top-K shape:

```json
[
  ["label", 0.73],
  ["another label", 0.61]
]
```

The first value is the label and the second is a route- and head-specific ranking score. The order matters. The decimal is not automatically a calibrated probability, and scores from different heads are not generally comparable.

## Inspecting a section

For each section:

1. Identify the exact field name.
2. Confirm whether the value is predicted, retrieved, composed, or caller-provided.
3. Read reliability and display-mode metadata.
4. Check whether support is attached to the section or to a specific item/pointer.
5. Read caveats, nulls, and withheld states.
6. Preserve route and release identity when copying the value.

For MoA, cognition, and side effects, [Section Inspector and Data Explorer](/start-using-neurolab/section-inspector-and-data-explorer.md) keeps these checks together. It can display the original rank and score, reliability and scope, report-returned evidence, and separately retrieved literature without modifying the report.

## Attaching report context to a question

The application lets you attach a section or selected context to Chat Report. Attached context narrows what the assistant can responsibly explain and helps it refer to exact fields.

Good questions are local and testable:

* “Why is this item ranked above the next one?”
* “Is the attached record a prediction or a sourced fact?”
* “Which support link points to this exact report item?”
* “What does the reliability display mode prevent me from concluding?”
* “Which route produced this brain channel?”

The assistant’s answer is a report explanation, not a new model evaluation or an independent literature review.

## Exports

Raw Payload supports section search, copy, JSON export of the full report, and JSON export of selected sections. The Knowledge Graph Explorer can export Graph JSON and visual renderings as SVG or PNG, depending on the active renderer.

Before sharing an export:

* retain model, KG, schema, and generated-at metadata;
* retain warnings and abstention state;
* avoid removing the route from a structure-derived result;
* retain source and license metadata for surfaced evidence;
* redact private or sensitive input; and
* do not present a graph image as a neural-network reasoning trace.

When comparing exports in a [Molecular Portfolio](/organize-and-compare-research/molecular-portfolios.md) or a future [Data Studio](/organize-and-compare-research/data-studio.md) workflow, keep model, facts-and-evidence, graph, schema, route, and generation identities visible. Compare unlike versions only when the difference itself is the subject of the analysis.

## Missing versus omitted

These states are different:

* **Empty:** the field exists but contains no items.
* **Null:** the contract represents the value as unavailable or not computed.
* **Omitted:** the field was not returned in this response shape or route.
* **Withheld / abstained:** the system deliberately suppressed a result under a reliability or applicability rule.
* **Not loaded:** a reference-only channel or source artifact was not supplied.

Preserving these distinctions is part of scientific honesty. Never fill them by extrapolating from an adjacent section.

Next: [How to Read a Neurolab Report](/start-using-neurolab/report-interpretation.md)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.neurolabai.io/start-using-neurolab/reports-and-raw-payload.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.