adapter

Definition

Plain language

A small piece of extra machinery bolted onto a model that can be trained while the original weights stay frozen.

Also called: adapters

Why it matters: Adapters make it cheap to specialize huge pretrained models without retraining or storing a separate full copy of weights per use case.

For example, you can attach a small adapter to a frozen base model and fine-tune only the adapter on a medical Q&A dataset.

Heard on the show

“The Model Context Protocol — think of it as a universal adapter, a USB-C port for AI tools.”

Mentioned in 10 episodes

1. 208
   The Blank Space in Your AI Approval Box That Isn't Empty
2. 192
   A 32B Open Model Matched Frontier Systems By Learning to Take Notes
3. 158
   How Floating-Point Rounding Lets a Model Tell Which Chip It's On — And Misbehave
4. 128
   How a Model Can Earn Full Reward and Still Resist Training
5. 114
   Agents That Rewrite Their Own Weights Instead of Just Taking Notes
6. 060
   When Splitting One Model Across Three Agents Doubles Its Accuracy
7. 040
   Two Frozen Models Learn to Whisper: Coupling Through Hidden States
8. 026
   What RL Actually Does to Language Models, at the Token Level
9. 025
   The Missing Gradient Term That Predicts Sycophancy in RLHF
10. 004
    The Sycophancy Circuit That Survives Alignment Training

Related concepts

Related terms