AI Glossary Term

What is Mid-training?

Definition

An intermediate training phase between pretraining and post-training, often used for capability or domain adjustments.

Related terms

Post-training
Training steps applied after pretraining, such as instruction tuning, preference optimization, and safety tuning.
Dataset
A collection of structured or unstructured examples used for training, validation, or testing.
Evaluation Set
A held-out dataset used to measure model quality after training.
Fine-Tuning
Continuing training on domain-specific data to adapt a pre-trained model to a specific task.
Pretraining
Initial large-scale model training on broad data before downstream adaptation.
Reward Model
A model that scores outputs based on preference signals, often used in RLHF pipelines.

Learn more in our free guides

AI TrainingLost in the Middle EffectPseudo-Labeling and Self-TrainingELECTRA Pretraining
Browse the full AI GlossaryExplore all free AI guides