naman-cen commited on
Commit
5ca70d9
·
verified ·
1 Parent(s): f368f9b

Simplify hero description

Browse files
Files changed (1) hide show
  1. app.py +4 -4
app.py CHANGED
@@ -909,10 +909,10 @@ THEME = gr.themes.Soft(primary_hue="violet", secondary_hue="indigo", neutral_hue
909
  HERO = """
910
  <div id="hero">
911
  <h1>🧠 SiriBench — On-Device iOS Agent Benchmarking</h1>
912
- <p>How we <b>author tasks</b> for an on-device iOS agent — Apple's Foundation Model driving the
913
- <b>real</b> Reminders, Calendar, Contacts & Messages and <b>capture every run as a verifiable trajectory</b>.
914
- Each task seeds a real iOS world, runs through a single Standard Agent with 11 tools, and is scored from the
915
- trajectory <i>plus</i> a re-read of the real OS store.</p>
916
 
917
  <div class="stats-grid">
918
  <div class="stat-card">
 
909
  HERO = """
910
  <div id="hero">
911
  <h1>🧠 SiriBench — On-Device iOS Agent Benchmarking</h1>
912
+ <p>This is a benchmark for evaluating <b>Apple's Foundation Model</b> the on-device model that
913
+ powers <b>Siri</b>. We created a set of real-world tasks across the actual iOS apps
914
+ (Reminders, Calendar, Contacts & Messages), let the model complete them on its own, and check
915
+ whether it actually got them right.</p>
916
 
917
  <div class="stats-grid">
918
  <div class="stat-card">