- formatting
- images
- links
- math
- code
- blockquotes
- external-services
•
•
•
•
•
•
-
The Generalization Imperative (4/4): Interpretability, Weak-to-Strong, and the Path Forward
Part 4 of 4 — Sections 10–12. Mechanistic interpretability of OOD generalization, weak-to-strong supervision and trust functions, and a synthesis across all 43 papers with references.
-
The Generalization Imperative (3/4): Causality, World Models, and Embodied AI
Part 3 of 4 — Sections 7–9. Causal representation learning for invariance, world models and physical law discovery, and embodied generalization tested in the real world.
-
The Generalization Imperative (2/4): Grokking, Meta-Learning, and OOD Generalization
Part 2 of 4 — Sections 4–6. Grokking dynamics and mode-hopping in real LLMs, meta-learning foundations, and OOD generalization through invariance and phase transitions.
-
The Generalization Imperative (1/4): From the Puzzle to Computation Theory
Part 1 of 4 — Sections 1–3. The puzzle of generalization, the architecture hypothesis, and computation-theoretic foundations.
-
Awesome Computational Imaging
Learning modern computational imaging through code and theory