Each note is an .mdx file in content/blog/. The file name becomes the URL, so this file (writing-research-notes.mdx) is served at /blog/writing-research-notes/.
Front matter
--- title: "Note title" publishedAt: "2026-10-11" updatedAt: "2026-10-12" # optional summary: "One-line summary shown in the list." tags: ["continual-learning", "paper-notes"] draft: true # optional, hides the note from the site ---
Math
Inline math such as and display math are rendered with KaTeX:
Code
import torch def fast_weight_update(W, k, v, lr=1.0): # Outer-product (Hebbian) write into the fast-weight matrix return W + lr * torch.outer(v, k)
Tables
| Setting | Value |
|---|---|
| Optimizer | AdamW |
| Learning rate | 1e-4 |
Delete this file once you add your first real note.