Clearer favicon/mark, bigger hero logo, drop blog em dashes

- Favicon + nav/footer mark + docs logo: white rounded tile + color octopus
  (octopus-icon.svg) so the mark stays legible at favicon/nav sizes; regenerate
  PNG fallbacks (favicon, apple-touch, icon-512) to match
- Hero: crop the octopus SVG to a tight portrait and widen the slot so the
  mascot reads larger
- Remove the unused raster mark (dps-mark.png)
- Replace em dashes with commas/colons/parentheses in the blog posts

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Author: SharonHart

astro.config.mjs

@@ -17,7 +17,7 @@ export default defineConfig({
       description:
         "Open-source building blocks for detecting, anonymizing, and protecting sensitive data.",
       logo: {
-        src: "./src/assets/dps-mark.png",
+        src: "./src/assets/octopus-icon.svg",
         replacesTitle: false,
       },
       favicon: "/favicon.png",
@@ -34,6 +34,7 @@ export default defineConfig({
         },
         { tag: "link", attrs: { rel: "preconnect", href: "https://fonts.googleapis.com" } },
         { tag: "link", attrs: { rel: "preconnect", href: "https://fonts.gstatic.com", crossorigin: true } },
+        { tag: "link", attrs: { rel: "icon", type: "image/svg+xml", href: "/octopus-icon.svg" } },
         { tag: "link", attrs: { rel: "apple-touch-icon", href: "/apple-touch-icon.png" } },
         {
           tag: "link",

public/apple-touch-icon.png

@@ -1,27 +1,27 @@
 ---
 title: "Evaluating PII Detection Models: A Field Guide"
-description: Why a single accuracy number can be misleading when you choose a PII detection model — and how to evaluate one properly.
+description: Why a single accuracy number can be misleading when you choose a PII detection model, and how to evaluate one properly.
 ---
 
 _By Omri Mendels_
 
 Suppose you're choosing a PII detection model for your organization or platform.
 You run two candidates on the same annotated test set: Model A reports an
 accuracy (say F2, which penalizes false negatives more) of 0.91, Model B reports
-0.78. You pick Model A — then deploy it, and privacy incidents keep happening.
+0.78. You pick Model A, then deploy it, and privacy incidents keep happening.
 
 What went wrong? Often, nothing with the model itself. The numbers were real;
 they just weren't measuring what you thought they were measuring.
 
 Evaluating a model for general Named Entity Recognition asks a single question:
-how well does it detect entity type X? **PII detection is different** — it serves
+how well does it detect entity type X? **PII detection is different**: it serves
 several distinct purposes at once (de-identification, redaction, discovery,
 compliance), and each demands a different way of measuring success. A single
 headline metric collapses those purposes into one number and hides the trade-offs
 that actually matter for privacy.
 
 This field guide walks through how to evaluate PII detection models in a way that
-reflects what you're really trying to achieve — the metrics that matter, the
+reflects what you're really trying to achieve: the metrics that matter, the
 pitfalls of borrowing NER evaluation wholesale, and how to reason about false
 negatives versus false positives when the cost of a miss is a leaked identity.
 

public/favicon.png

@@ -7,9 +7,9 @@ News, releases, and notes from the Data Privacy Stack community.
 
 ## Posts
 
-- [Evaluating PII Detection Models: A Field Guide](/blog/evaluating-pii-detection-models/)
-  — why a single accuracy number can mislead when choosing a PII detection model.
-- [Welcome to the Data Privacy Stack](/blog/welcome/) — the community home for
+- [Evaluating PII Detection Models: A Field Guide](/blog/evaluating-pii-detection-models/):
+  why a single accuracy number can mislead when choosing a PII detection model.
+- [Welcome to the Data Privacy Stack](/blog/welcome/): the community home for
   open, transparent tools for private data handling.
 
 _Want to write a post? See how to [get involved](/#community)._