Validation

Model performance and validation evidence

This page summarizes the published evaluation so buyers can understand the performance level behind the docs bundle.

Decision Accuracy

0.8497

Decision Precision

0.8082

Decision Recall

0.8431

Decision F1

0.8230

Published run context

Field	Value
Initial evaluation period	08/2025
Published re-evaluation run	2025-10-28 (UTC log)
Model family	bert-base-uncased
Training dataset size	181,000 samples
Validation processing	711 batches
Thresholding	0.1, 0.1, 0.1

Additional evaluation notes

Metrics were captured from a reproducible internal validation workflow.
The published screenshot is intended as evidence, not as a marketing graphic.
The page should remain concise and focused on the latest approved public run.

These metrics reflect the current published evaluation snapshot.