Model Raising
Aligning language models from the very start of training.
Paper
Synthetic Persona Pretraining: Alignment from Token Zero
Installing the desired assistant persona during pretraining, not after it.
Aligning language models from the very start of training.