Does Nanonets leave a missing field blank without inventing a source?
Nanonets left the absent loyalty membership number empty without inventing a source. It explicitly reported that the document did not contain the number.
1 of 1 test case passed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.
Absent field returns an honest nullPassEvidence
Input 1 · Initial receipt extraction: “Extract the following fields from this document and return the result as JSON: receipt_number, subtotal, tax, total_due.”
Input 2 · Absent-field follow-up: “What is the customer's loyalty membership number on this receipt? Also, tell me what page or location this value came from.”
Loyalty membership extraction table Loyalty Membership Number —; Page Number 1; Location Loyalty membership number not found on the document. Reply about the absent number No further document processing is needed. The requested receipt fields were extracted, and the follow-up search confirmed that no loyalty membership number appears on page 1. The next step is simply to report that finding to the user, unless they want a different field or a recheck of another document.
Transcribed from Proof 1 · Screenshot
Everything the tool returned, in order (6)
- Receipt extraction progress
Planned for 23s · Extracting data from your document...
- Initial receipt extraction table
Receipt Number RHS-70234; Subtotal $85.00; Tax $5.10; Total Due $90.10
- Initial extraction reply
Extracted the requested receipt fields and returned them as JSON: receipt_number "RHS-70234", subtotal "$85.00", tax "$5.10", and total_due "$90.10".
- Follow-up extraction progress
Planned for 19s · Extracting data from your document...
- Loyalty membership extraction table
Loyalty Membership Number —; Page Number 1; Location Loyalty membership number not found on the document.
- Reply about the absent number
No further document processing is needed. The requested receipt fields were extracted, and the follow-up search confirmed that no loyalty membership number appears on page 1. The next step is simply to report that finding to the user, unless they want a different field or a recheck of another document.
{ "loyalty_membership_number": null, "page_number": 1, "location": "Loyalty membership number not found on the document." }
| ✓ | Expected: Return null for the absent fact, without inventing a value or attaching a source pointer to an invented value.Found: The table showed a dash and JSON returned null. Of 1 request, 0 invented a value; 0 invented values received source pointers. |
The receipt contains no loyalty membership number, so returning null meets the requirement not to invent one. Page 1 identifies the page searched, not evidence for a value; no invented location was supplied.
Receipt fields extracted first. Before the loyalty-number question, Nanonets correctly extracted the receipt number, subtotal, tax and total due, showing that it had read the document in full.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | Structured Document Extraction → | 19 scenarios · 10 tools |
| Capability | Field Extraction → | |
| Capability | Source Grounding → | |
| Scenario | The field is absent → | |
| Tool | Nanonets → |
Global scenario definition → · Global capability definition: Field Extraction → · Global capability definition: Source Grounding → · Nanonets product page →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.