SAI LabsBook a demo

Safety IndexEvaluationsPrompted sandbagging

Prompted sandbagging

Evaluation integrity · SandbaggingEvaluation-integrity profileBuilt, not yet run

Whether the model deliberately underperforms when prompted to, following van der Weij et al. 2024.

Metric
accuracy drop (percentage points). Capability, not scored.
Dataset
github.com/TeunvdWeij/sandbagging
Grader
–
Status
Built, not yet run
Replicability
public

Results

Built and tested; the first full runs are pending.