Six-Dimensional Taxonomy of Post-Training Adaptation

Operational AI RMF change tracking needs precise vocabulary for post-training adaptations applied to deployed models.

Taxonomy vocabulary is explicitly pitched for technical documentation, model-change tracking, and governance.

Taxonomy aims to standardize how trained-model modifications are described for documentation and governance.

Shared governance need to document how deployed models are modified after training.

A shared vocabulary for how models were adapted is a prerequisite for assigning institutional accountability after autonomous actions.

A shared vocabulary for how models are adapted can make downstream risk taxonomies more decision-relevant and auditable.

Shared vocabulary for how models were adapted improves documentation, change tracking, and evaluability for governors.