A measured baseline
Task-specific results with retained model outputs and explicit grading rules.
Verified data. Inspectable evidence.
Find where your AI is wrong. Build the data to improve it.
Groundtruth Data helps AI teams evaluate factual answers against authoritative sources, understand failures, and build targeted remediation data.
How it works
A clear path from evaluation to better data.
Start with one important task. Keep the evidence, the correction, and the later validation connected.
Define the task
Choose the question your system needs to answer and the sources that can establish correctness.
Measure the baseline
Run a source-backed evaluation, retain model outputs, and inspect the evidence behind each grade.
Target the failures
Use verified corrections to scope remediation data. Your team controls changes to its model or system.
Evaluate the change
Test on separate cases. Report what changed, including regressions and unresolved errors.
What you get
More than a score.
Failures with receipts
Source references, verified answers, and examples your team can inspect.
A focused next step
Targeted remediation and a separate validation plan, scoped to what the evidence supports.
Packaged dataset
500-case Drug Indication Proof Evaluation
A historical MED-RT source-grounding evaluation: 500 unique questions, verified reference answers, source provenance, and recorded model grades.
- JSONL evaluation and provenance files
- Historical model responses and grades
- SHA-256 manifest and internal-use license
This is separate from the retired 338-row diagnostic export. It is not clinical advice, an independently blinded holdout, or evidence of model improvement.
Read the evidence
What 500 SEC filing questions revealed.
Our recorded proof evaluation returned 136 correct answers out of 500. Read the task definition, failure patterns, and limits behind that result.
Read the study →Work with Groundtruth
Start with your team's task.
For a custom pilot, we agree on sources, scope, price, and turnaround before work begins. Your team retains control of model and system changes.
Get started