Training-Time-Only Model Logic

Kind: anti-pattern

Layer: Correctness Core

Record: lexicon:training-time-only-model-logic

Canonical: Lexicon

Building logic that exists only during training, with no counterpart to serve predictions at inference.

Listed in Lexicon terms, after Opaque Ungoverned Model Use and before Unapproved Model Deployment.

Category

• Model Architecture

Negated by

• Model Inference

Linked from

• Model Architecture