Installs as a private draft. Your edits and self-improvement runs do not change the published bundle.
Detect when RL training narrows capability (great on trained tasks, worse on everything else)
Detect when RL training narrows capability (great on trained tasks, worse on everything else)
RL can make models worse at things you didn't train on
Generalization & Continual Learning > Generalization
--- name: "Overfitting Detection For Rl" version: "1.0.0" description: "Detect when RL training narrows capability (great on trained tasks, worse on everything else)" tags: - rl-skills-universe - "generalization-continual-learning" - "generalization" trigger: "overfitting-detection-for-rl" --- # Overfitting Detection For Rl Detect when RL training narrows capability (great on trained tasks, worse on everything else) ## Why It Matters RL can make models worse at things you didn't train on ## Domain **Generalization & Continual Learning** > Generalization