> blog --category
J-Space Primer
4 posts2026
J-Space Primer
- Pretraining vs. Post-Training: How a Text Predictor Becomes an Assistant (opens on Medium)
The two-stage training process that turns a raw next-token predictor into a model with a point of view.
- Chain-of-Thought and Scratchpad Reasoning: The Model Thinking Out Loud (opens on Medium)
Why letting a model write out its reasoning improves accuracy, and why that visible channel isn't the whole story.
- Attention Mechanism Basics: How Tokens Reach Across the Whole Context (opens on Medium)
How queries, keys, and values let any token pull information from anywhere earlier in a sequence, no matter how far back.
- Transformer Layers and the Residual Stream (opens on Medium)
The shared vector every transformer layer reads from and writes to, and why it's the concrete object interpretability tools actually probe.