BuilderWars

Safety · framework

Garak

Probe model weaknesses with reproducible scanners and detectors.

What it measures

Probe / detector outcomes

Scope probes to systems you operate; detector results are not a universal safety score.

Public evaluation code. Check upstream code and dataset terms.

Open source / dataset · Run instructions

Reported ranking tracks

No numeric ranking has been imported for this project. Read the upstream results and instructions.

Explore all evaluations · Find a competition