How VeriLM grades.
The point of a Deep Research report is to give you claims you can verify against a source — so you know you aren't reading a hallucination. VeriLM grades every report against that bar, sentence by sentence.
The grading process
Decompose
Every sentence that makes a factual statement is broken into individual claims. This happens whether or not the sentence is cited — uncited claims don't get to skip the grading.
Assign sources
Each sentence with factual claims gets its cited sources attached. A single sentence can pull from multiple sources.
Grade each claim
Every claim is checked against each of its assigned sources individually. If a claim fails against every individual source, a synthesis pass runs — checking the claim against all referenced sources together — before the claim is marked unsupported.
Roll up the verdict
A sentence is verified only if every claim in it is verified. One unsupported claim drags the whole sentence into the unsupported bucket.
The six verdicts
This is how Deep Research grades, under the All Facts contract — every factual sentence has to trace to a source, and one that doesn't is a failure, not an omission.
| Verdict | Definition |
|---|---|
| Verified | Every claim in the sentence checks out against at least one of its cited sources. |
| Unsupported — Wrong | A cited source speaks to the claim but contradicts it. Example: the report says "Raleigh gets 213 sunny days per year," but the cited source says 200. |
| Unsupported — Missing | The cited source doesn't establish the claim — either it's silent on it, or the claim's core holds but it adds a specific the source never states. Example: the source supports the payback estimate, but the sentence also attributes it to "a 2024 NREL study" the source never mentions. |
| Unsupported — Uncited | The sentence makes a factual claim but has no source attached. |
| Unresolved — Unverified | Grading couldn't complete — either the cited source couldn't be retrieved, or our own grader failed to return a usable verdict. Neither is a fault of the report, so neither is counted against it. |
| Unresolved — Narrative | The sentence doesn't make a factual claim. Ungraded by definition. |
Outside Deep Research, VeriLM also offers a lighter Key Facts contract, where a model may draw on its own knowledge as long as it sources the load-bearing facts. There an uncited claim is flagged rather than counted as a failure, so it moves out of Unsupported and out of the accuracy denominator. Same six verdicts, one different grouping.
The accuracy formula
Unresolved sentences — narrative and unverifiable — are excluded from both numerator and denominator. Narrative doesn't make factual claims, so there's nothing to grade. Unverifiable means the cited source couldn't be retrieved — an operational miss, not a content failure.
Want to see VeriLM Deep Research in action? Request beta access below.