RootLens

Benchmark runs

Each run replays the incident scenarios against hidden ground truth and scores the result. Accuracy varies with the model and hardware — these are measured, not targets.

Loading…