masks post: size diagram figure to 720 so it reads clearly

Author: HumbleBee14

src/content/posts/masks-in-transformers/index.mdx

@@ -28,7 +28,7 @@ Strip away the jargon and every mask is the same primitive:
 
 Four purposes, four families:
 
-<Figure caption="One primitive, four purposes. Every 'mask' in deep learning is a 0/1 gate that says 'ignore this' — the families differ only in what they ignore, and why.">
+<Figure caption="One primitive, four purposes. Every 'mask' in deep learning is a 0/1 gate that says 'ignore this' — the families differ only in what they ignore, and why." width={720}>
 <svg viewBox="0 0 720 224" role="img" aria-label="A mask is a 0/1 gate; it splits into four families — SEE (attention masks), SCORE (loss masks), HIDE (objective masks), and DROP (regularization masks)." style="width:100%;max-width:720px;height:auto;color:currentColor;font-family:var(--font-mono);">
   <g fill="none" stroke="currentColor" stroke-width="1.3">
     <rect x="270" y="12" width="180" height="48" rx="8" opacity="0.6"/>