Selection Scope
Only completed simulations with available analyses can be selected for cross-simulation benchmarking.
Benchmark Comparison
Compare semantic network metrics across completed crisis simulations.
0 Simulations0 Selected
Ready for Benchmarking
Metrics Benchmarking
Select two or more simulations from the sidebar to calculate aggregated metrics across semantic networks.
MethodologyBenchmark MethodologyCommon criteria for selecting completed audits, aggregating model metrics, and comparing semantic network behavior.
View methodologyHide
Common criteria for selecting completed audits, aggregating model metrics, and comparing semantic network behavior.
Aggregation
Metrics are grouped by AI model across the selected simulations and summarized with mean and standard deviation.
Topology Comparison
Semantic network metrics, CASA adherence, null-model baselines, and assortativity indicators are compared across selected audits.
Export
The exported CSV preserves model-level means and standard deviations for every benchmark metric shown in the matrix.
Interpretation: Use this view to compare model behavior across simulations, not to inspect a single audit in depth.