Welcome to the blog. I’ll use this space to document learning notes on LLMs, post-training, and personalized AI. This first post doubles as a rendering test.

Headings & text

Regular paragraphs, bold, italic, inline code, and links.

Blockquotes look like this — handy for highlighting a key idea.

Code

def reward(model, prompt, response):
    """A toy reward function."""
    return model.score(prompt, response)

Math

Inline math like $f(x) = \langle w, x \rangle + b$ renders with KaTeX, and display math too:

$$ \mathcal{L}(\theta) = -\mathbb{E}_{(x, y)\sim \mathcal{D}}\big[\log \pi_\theta(y \mid x)\big] $$

Lists

  1. Write a draft (hugo new posts/my-post.md).
  2. Set draft: false when ready.
  3. Push — the site rebuilds and deploys automatically.

That’s it. 🐱