Model-card index¶

This index records the public model documentation for procurement, security, and AI-governance review.

Model documentation

AI transparency, Confidence & evidence, and Known limitations document model behavior, confidence, fallback handling, and review controls.

Customer-facing artifacts for each model:

Model	Purpose	Status
Stage B attribution model	Scores edge-existence evidence for ambiguous residuals left by deterministic resolution, within the asynchronous processor’s 20 ms Stage B timeout.	Standalone public model card: not published.

Customer-facing artifacts for the Stage B attribution model: AI transparency; Confidence & evidence.

Card coverage¶

A model card is listed as published only when it covers:

model purpose and non-purpose;
training and evaluation data boundary;
confidence and calibration semantics (including the 0.95 operational cap);
known limitations and failure modes, including the deterministic heuristic-baseline fallback;
human review, promotion, and rollback controls (promotion is a human decision; canary rollback never touches your traffic);
model version, publication date, and evidence owner;
customer-facing change history.

Sign in to send feedback. Opens in a new tab.

Copy page details

Docs feedback: Model-card index
Page: https://docs.venturi.systems/engine/model-card-index/