update the demo description

Author: Shuyang-Liu

.gitattributes

@@ -0,0 +1 @@
+*.mp4 filter=lfs diff=lfs merge=lfs -text

index.html

@@ -109,13 +109,19 @@ <h1 class="title is-1 publication-title">
 
 
 <section class="section">
-  <div class="container" style="margin-top: -2rem;">
-    <div class="columns is-centered">
-      <div class="column is-four-fifths has-text-centered">
-        <div class="content">
-          <img src="./static/images/html_demo.png" alt="Graphectory demo" style="width: 100%; border-radius: 12px;"/>
-          <p style="font-size: 1.05rem; margin-top: 1rem;">
-            Interactive visualization of Graphectory for process-centric analysis of agent trajectories.
+  <div class="container is-max-desktop">
+    <div class="columns is-centered has-text-centered">
+      <div class="column is-four-fifths">
+        <h2 class="title is-3">Abstract</h2>
+        <div class="content has-text-justified" style="font-size: 1.12rem;">
+          <p>
+            Agentic systems are modern software systems: they consist of orchestrated modules, expose interfaces, and are deployed in software pipelines. Unlike conventional programs, their execution trajectories are inherently stochastic and adaptive to the problems they solve. Evaluation of such systems is often outcome-centric, focusing only on whether the final task succeeds or fails. This narrow view overlooks how agents reason, plan, act, and revise their strategies during execution.
+          </p>
+          <p>
+            We introduce <b>Graphectory</b>, a graph-based representation that systematically encodes the temporal and semantic relations in agent trajectories, enabling process-centric metrics and analyses of workflow quality. Using Graphectory, we analyze 4,000 trajectories from two leading software engineering agents, <b>SWE-agent</b> and <b>OpenHands</b>, paired with four backbone large language models on SWE-bench Verified.
+          </p>
+          <p>
+            Our fully automated analyses reveal that stronger models and richer prompts lead to more complex yet more coherent workflows, while unsuccessful trajectories tend to exhibit chaotic, repetitive, or backtracking behaviors. We further develop a real-time monitoring framework that constructs and analyzes <b>Graphectory</b> and <b>Langutory</b> online to detect process-level issues, notify the agent with diagnostic feedback, and roll back execution when appropriate. Experiments show that online monitoring with intervention improves resolution rates by 6.9%–23.5% on problematic instances while significantly shortening trajectories with near-zero overhead.
           </p>
         </div>
       </div>
@@ -126,26 +132,37 @@ <h1 class="title is-1 publication-title">
 
 <section class="section">
   <div class="container is-max-desktop">
-    <div class="columns is-centered has-text-centered">
+    <div class="columns is-centered">
       <div class="column is-four-fifths">
-        <h2 class="title is-3">Abstract</h2>
-        <div class="content has-text-justified" style="font-size: 1.12rem;">
+
+        <h2 class="title is-3 has-text-centered">Interactive Demo</h2>
+
+        <div class="content has-text-justified" style="font-size: 1.12rem; margin-bottom: 1.5rem;">
           <p>
-            Agentic systems are modern software systems: they consist of orchestrated modules, expose interfaces, and are deployed in software pipelines. Unlike conventional programs, their execution trajectories are inherently stochastic and adaptive to the problems they solve. Evaluation of such systems is often outcome-centric, focusing only on whether the final task succeeds or fails. This narrow view overlooks how agents reason, plan, act, and revise their strategies during execution.
+            <b>Graphectory</b> provides interactive visualization and process-centric analysis of agent trajectories, enabling users to inspect how software engineering agents reason, explore, patch, and validate over time.
           </p>
           <p>
-            We introduce <b>Graphectory</b>, a graph-based representation that systematically encodes the temporal and semantic relations in agent trajectories, enabling process-centric metrics and analyses of workflow quality. Using Graphectory, we analyze 4,000 trajectories from two leading software engineering agents, <b>SWE-agent</b> and <b>OpenHands</b>, paired with four backbone large language models on SWE-bench Verified.
+            Our current implementation supports trajectories from <b>SWE-agent</b>, <b>OpenHands</b>, and <b>mini-swe-agent</b>. The framework is designed to be extensible, allowing users to adapt it with minimal effort to other agentic systems.
           </p>
+        </div>
+
+        <div class="has-text-centered">
+          <img src="./static/images/html_demo.png"
+               alt="Interactive visualization of Graphectory"
+               style="width: 100%; border-radius: 12px;" />
+        </div>
+
+        <div class="content has-text-centered" style="margin-top: 1rem; font-size: 1.02rem;">
           <p>
-            Our fully automated analyses reveal that stronger models and richer prompts lead to more complex yet more coherent workflows, while unsuccessful trajectories tend to exhibit chaotic, repetitive, or backtracking behaviors. We further develop a real-time monitoring framework that constructs and analyzes <b>Graphectory</b> and <b>Langutory</b> online to detect process-level issues, notify the agent with diagnostic feedback, and roll back execution when appropriate. Experiments show that online monitoring with intervention improves resolution rates by 6.9%–23.5% on problematic instances while significantly shortening trajectories with near-zero overhead.
+            Interactive visualization of Graphectory for trajectory inspection and process-centric analysis.
           </p>
         </div>
+
       </div>
     </div>
   </div>
 </section>
 
-
 <section class="section" id="BibTeX">
   <div class="container is-max-desktop content">
     <h2 class="title is-3 has-text-centered">BibTeX</h2>