ClarusC64/cascade-multi-ai-model-release-misuse-policy-response-v0.1

What this repo does

This dataset tests whether a model can detect an AI governance cascade.

You give the system:

• —model release conditions
• —deployment scale
• —misuse pressure signals
• —detection and policy lag
• —trust and coupling signals

You ask it to:

• —predict whether the scenario crosses into a cascade event

Core quad

This cascade family can include more nodes.

This repo anchors a core quad inside the wider cascade:

• —releasecontrolsstrength
• —misuseincidentsrate
• —policyresponselag_days
• —publictrustindex

These four variables form a stable cascade signature across many AI deployment narratives.

Prediction target

Target column:

• —labelcascadeevent

Meaning:

• —0 = strain stays containable
• —1 = cascade event forms and propagates across governance, media, and platform response loops

Row structure

Each row is a scenario snapshot.

Key columns:

• —modelcapabilitylevel
• —releasecontrolsstrength
• —redteamcoverage
• —deployment_scale
• —misuseincidentsrate
• —detectionlatencydays
• —policyresponselag_days
• —commsclarityscore
• —publictrustindex
• —couplingmediaregulator
• —couplingplatformgovernment
• —buffermitigationcapacity
• —cascadeseverityscore

Files

• —data/train.csv 10-line sample with labels

• —data/tester.csv 10-line sample with labels

• —scorer.py Binary classification metrics and confusion matrix

Evaluation

Run local scoring by comparing a gold CSV to a prediction CSV that contains label_cascade_event.

Example:

• —gold = data/tester.csv
• —pred = your_predictions.csv

Command:

python scorer.py --gold data/tester.csv --pred your_predictions.csv

Metrics:

• —accuracy
• —precision
• —recall
• —f1
• —confusion matrix

License

MIT

This dataset identifies a measurable coupling pattern associated with systemic instability. The sample demonstrates the geometry. Production-scale data determines operational exposure.

What Production Deployment Enables • 50K–1M row datasets calibrated to real operational patterns • Pair, triadic, and quad coupling analysis • Real-time coherence monitoring • Early warning before cascade events • Collapse surface and recovery window modeling • Integration and implementation support Small samples reveal structure. Scale reveals consequence.

Enterprise & Research Collaboration Clarus develops production-scale coherence monitoring infrastructure for critical systems across healthcare, finance, infrastructure, and regulatory domains. For dataset expansion, custom coherence scorers, or deployment architecture: team@clarusinvariant.com

Instability is detectable. Governance determines whether it propagates.