freeze

Definition

To stop a model's weights from changing during further training.

To hold a set of parameters fixed during optimization, typically while training only adapters, communication layers, or downstream modules.

Also called: frozen, frozen weights

Mentioned in 17 episodes

1. 088
   Two Levers for Self-Improving AI: When Rewriting Code Isn't Enough
2. 086
   Why Frozen-Weight Agents Still Get Worse Over Time
3. 083
   Training the Translator: How a Small Communication Model Lets Agent Teams Outperform Themselves
4. 073
   When Three LLMs Talk to Each Other, Their Ideas Quietly Stop Moving
5. 071
   When the Model Is Fine and the Plumbing Is Broken: Fixing Agents at the Interface
6. 069
   When Smarter Models Forecast Worse: The Hidden Failure Mode in LLM Predictions
7. 068
   The OS Trick That Makes Tree Search Practical for Coding Agents
8. 060
   When Splitting One Model Across Three Agents Doubles Its Accuracy
9. 048
   How a 30B Open Model Reached Olympiad Gold With the Right Recipe
10. 040
    Two Frozen Models Learn to Whisper: Coupling Through Hidden States
11. 032
    A Sticky-Note for Every Layer: Letting Transformers Remember What They Were Just Thinking
12. 028
    Teaching a Model to Hire Copies of Itself: Recursive Agent Optimization
13. 026
    What RL Actually Does to Language Models, at the Token Level
14. 025
    The Missing Gradient Term That Predicts Sycophancy in RLHF
15. 019
    When the Best Reward Model Trains the Worst Policy: Inside EvoLM
16. 018
    Language Models Compute the Rational Move, Then Override It
17. 016
    Why Your Coding Agent Stalls While the GPU Runs Hot

Related concepts