Reformulate subtitle

AGENTS.md

@@ -11,7 +11,7 @@ run no compute, so they are free; do not switch it to Gradio or Docker (those ne
   Elo figure above the table; Performance Trade-offs: Pareto front and complete results; Stability;
   Validity),
   all user-facing text (each explanation sits next to the figure or table it describes; there is no
-  About tab; the Paper and GitHub links sit in the intro sentence), the citation (always visible
+  About tab; the Paper and GitHub links are the header buttons), the citation (always visible
   in the footer) and `<meta name="selectarena-version">` (not displayed; there is no version
   history, the commit messages are the changelog).
 - `assets/app.js`: plain JavaScript, no build step. Fetches `data/meta.json`, renders the selector

index.html

@@ -4,7 +4,7 @@
   <meta charset="utf-8">
   <meta name="viewport" content="width=device-width, initial-scale=1">
   <!-- Bumped with every data refresh (shown in the header). -->
-  <meta name="selectarena-version" content="0.8.6">
+  <meta name="selectarena-version" content="0.8.7">
   <title>SelectArena Leaderboard</title>
   <meta name="description" content="Leaderboard of feature selection methods on real-world tabular data.">
   <link rel="icon" type="image/svg+xml" href="assets/logo.svg">
@@ -39,7 +39,7 @@
   </header>
 
   <p class="intro">
-    SelectArena is a <b>living</b> benchmark. We compare predictive performance, stability and validity of feature selection methods. To add your feature selection method to the benchmark, see our <a href="https://github.com/schaeferbasti/SelectArena/tree/main#add-your-method-to-selectarena" target="_blank" rel="noopener">GitHub</a>.
+    Which feature selection method should I use? SelectArena compares them on predictive performance, stability and validity. It is a living benchmark: more method will be added.
   </p>
 
   <div class="tabs" role="tablist" aria-label="Leaderboard sections">