Caloogy Article

LLM Frontier

Interventional Training: Letting Models Conduct Their Own Experiments

LLM Frontier · Sep 19, 2026 · 4 min

Beyond Tokens: Thinking in Latent Space

LLM Frontier · Jun 13, 2026 · 4 min

Test-Time Training: Models That Update During Inference

LLM Frontier · May 20, 2026 · 3 min

Beyond Verifiable Rewards: How to Do Reinforcement Learning in Domains Without Standard Answers

LLM Frontier · May 10, 2026 · 4 min

Free Energy Principle as a Training Objective: Replacing Reward Maximization with "Minimizing Surprise"

LLM Frontier · Dec 11, 2025 · 4 min

Post-training with High-Dimensional World Models: From "Predicting the Next Token" to "Predicting the Next State"

LLM Frontier · Oct 19, 2025 · 5 min