Paper summary
Paper summary: On the Scaling of PEFT: Towards Million Personal Models of Trillion Parameters
This position-style study argues that small adapters on top of shared foundation models can serve as persistent personal state (preferences, skills, tool habits, memory-like updates), organised along Scale Up, Scale Down and Scale Out axes.
- The base model gives shared competence; each adapter carries one instance's personal behaviour.
- PEFT is reframed from a budget substitute for full fine-tuning to a substrate for persistent personal models.