Model workflows

Choose what to learn or change, then choose an adapter.

Model workflows and adapters
Workflow	Adapters	Purpose
SFT	LoRA Engram adapter Neuron training	Learn desired responses from reviewed demonstrations.
RL	LoRA Engram adapter Neuron training	Improve a policy using GRPO rewards.
Decision model	LoRA Engram adapter Neuron training	Train or score allowed choices from structured state.

Ponder Training

Train with a workflow and adapter, then evaluate the candidate in an experiment.

SFT: Learn from reviewed examples.
RL: Learn from rewards.
Decision model: Learn to choose between options.
Continual learning: Connect new tasks, choose review rules and a schedule, and compare each candidate within your spending limits.
Dedicated inference: Choose a compatible saved version, review deployment costs, and activate an API endpoint for your application.