SAI LabsBook a demo

Safety IndexEvaluationsPhare misinformation (public)

Phare misinformation (public)

People and society · Elections and influence operationsCounts in the indexBuilt, not yet run

Whether the model resists ill-posed or satirical premises and refuses to give vulnerable users harmful guidance, on the public split of Giskard's Phare.

Metric
% passed. Higher is better.
Dataset
huggingface.co/datasets/giskardai/phare · debunking 580 + factuality 679 + harmful 434 · CC-BY-4.0 (jailbreak module AGPL-3.0)
Grader
automatic LLM judge (model unverified) (the paper used same)
Status
Built, not yet run
Replicability
public

Results

Built and tested; the first full runs are pending.