link test

Author: hasithv

content/posts/25-09-29-nanodiffgpt/index.md

@@ -255,8 +255,10 @@ It was a fun little experiment to convert nanoGPT into a diffusion LLM. I've bee
 
 [1] GitHub repo - nanoDiffGPT: https://github.com/hasithv/nanoDiffGPT
 
-[2] LLaDa paper: https://arxiv.org/abs/2502.09992
+[2] Diffusion notes: [{{< ref "../FlowDiffusion/FlowDiff.md" >}}]({{< relref "../FlowDiffusion/FlowDiff.md" >}})
 
-[3] Block diffusion paper: https://arxiv.org/abs/2503.09573
+[3] LLaDa paper: https://arxiv.org/abs/2502.09992
 
-[4] MMaDa paper: https://arxiv.org/abs/2505.15809
+[4] Block diffusion paper: https://arxiv.org/abs/2503.09573
+
+[5] MMaDa paper: https://arxiv.org/abs/2505.15809

public/archives/index.html

@@ -206,10 +206,10 @@ <h3 class="archive-month-header" id="2025-September">
     </h3>
     <div class="archive-posts">
       <div class="archive-entry">
-        <h3 class="archive-entry-title entry-hint-parent">Making Nano-GPT a Diffusion LLM
+        <h3 class="archive-entry-title entry-hint-parent">Hacking Nano-GPT into a Diffusion LLM
         </h3>
         <div class="archive-meta"><span title='2025-09-29 19:45:01 -0500 CDT'>September 29, 2025</span>&nbsp;·&nbsp;10 min</div>
-        <a class="entry-link" aria-label="post link to Making Nano-GPT a Diffusion LLM" href="https://hasithv.github.io/posts/25-09-29-nanodiffgpt/"></a>
+        <a class="entry-link" aria-label="post link to Hacking Nano-GPT into a Diffusion LLM" href="https://hasithv.github.io/posts/25-09-29-nanodiffgpt/"></a>
       </div>
       <div class="archive-entry">
         <h3 class="archive-entry-title entry-hint-parent">Lecture 2 - Constructing the Training Target

public/categories/projects/index.html

@@ -152,7 +152,7 @@ <h1>
 
 <article class="post-entry tag-entry"> 
   <header class="entry-header">
-    <h2 class="entry-hint-parent">Making Nano-GPT a Diffusion LLM
+    <h2 class="entry-hint-parent">Hacking Nano-GPT into a Diffusion LLM
     </h2>
   </header>
   <div class="entry-content">
@@ -161,7 +161,7 @@ <h2 class="entry-hint-parent">Making Nano-GPT a Diffusion LLM
 ...</p>
   </div>
   <footer class="entry-footer"><span title='2025-09-29 19:45:01 -0500 CDT'>September 29, 2025</span>&nbsp;·&nbsp;10 min</footer>
-  <a class="entry-link" aria-label="post link to Making Nano-GPT a Diffusion LLM" href="https://hasithv.github.io/posts/25-09-29-nanodiffgpt/"></a>
+  <a class="entry-link" aria-label="post link to Hacking Nano-GPT into a Diffusion LLM" href="https://hasithv.github.io/posts/25-09-29-nanodiffgpt/"></a>
 </article>
 
 <article class="post-entry tag-entry"> 

public/categories/projects/index.xml

@@ -9,7 +9,7 @@
     <lastBuildDate>Mon, 29 Sep 2025 19:45:01 -0500</lastBuildDate>
     <atom:link href="https://hasithv.github.io/categories/projects/index.xml" rel="self" type="application/rss+xml" />
     <item>
-      <title>Making Nano-GPT a Diffusion LLM</title>
+      <title>Hacking Nano-GPT into a Diffusion LLM</title>
       <link>https://hasithv.github.io/posts/25-09-29-nanodiffgpt/</link>
       <pubDate>Mon, 29 Sep 2025 19:45:01 -0500</pubDate>
       <guid>https://hasithv.github.io/posts/25-09-29-nanodiffgpt/</guid>
@@ -246,9 +246,10 @@ Hasting in a roseman.
 <p>It was a fun little experiment to convert nanoGPT into a diffusion LLM. I&rsquo;ve been thinking of other experiments on how to hijack architectures to make them do what I want, and so doing this was a good proof of concept as to how well it would work. Additionally, I got to explore diffusion LLMs, an area which I think holds tons of promise.</p>
 <h2 id="references">References</h2>
 <p>[1] GitHub repo - nanoDiffGPT: <a href="https://github.com/hasithv/nanoDiffGPT">https://github.com/hasithv/nanoDiffGPT</a></p>
-<p>[2] LLaDa paper: <a href="https://arxiv.org/abs/2502.09992">https://arxiv.org/abs/2502.09992</a></p>
-<p>[3] Block diffusion paper: <a href="https://arxiv.org/abs/2503.09573">https://arxiv.org/abs/2503.09573</a></p>
-<p>[4] MMaDa paper: <a href="https://arxiv.org/abs/2505.15809">https://arxiv.org/abs/2505.15809</a></p>
+<p>[2] Diffusion notes: <a href="/posts/flowdiffusion/flowdiff/">https://hasithv.github.io/posts/flowdiffusion/flowdiff/</a></p>
+<p>[3] LLaDa paper: <a href="https://arxiv.org/abs/2502.09992">https://arxiv.org/abs/2502.09992</a></p>
+<p>[4] Block diffusion paper: <a href="https://arxiv.org/abs/2503.09573">https://arxiv.org/abs/2503.09573</a></p>
+<p>[5] MMaDa paper: <a href="https://arxiv.org/abs/2505.15809">https://arxiv.org/abs/2505.15809</a></p>
 ]]></content:encoded>
     </item>
     <item>

public/index.html

@@ -219,7 +219,7 @@ <h1>Hello!</h1>
 
 <article class="post-entry"> 
   <header class="entry-header">
-    <h2 class="entry-hint-parent">Making Nano-GPT a Diffusion LLM
+    <h2 class="entry-hint-parent">Hacking Nano-GPT into a Diffusion LLM
     </h2>
   </header>
   <div class="entry-content">
@@ -228,7 +228,7 @@ <h2 class="entry-hint-parent">Making Nano-GPT a Diffusion LLM
 ...</p>
   </div>
   <footer class="entry-footer"><span title='2025-09-29 19:45:01 -0500 CDT'>September 29, 2025</span>&nbsp;·&nbsp;10 min</footer>
-  <a class="entry-link" aria-label="post link to Making Nano-GPT a Diffusion LLM" href="https://hasithv.github.io/posts/25-09-29-nanodiffgpt/"></a>
+  <a class="entry-link" aria-label="post link to Hacking Nano-GPT into a Diffusion LLM" href="https://hasithv.github.io/posts/25-09-29-nanodiffgpt/"></a>
 </article>
 
 <article class="post-entry"> 

public/index.json

@@ -9,7 +9,7 @@
     <lastBuildDate>Mon, 29 Sep 2025 19:45:01 -0500</lastBuildDate>
     <atom:link href="https://hasithv.github.io/index.xml" rel="self" type="application/rss+xml" />
     <item>
-      <title>Making Nano-GPT a Diffusion LLM</title>
+      <title>Hacking Nano-GPT into a Diffusion LLM</title>
       <link>https://hasithv.github.io/posts/25-09-29-nanodiffgpt/</link>
       <pubDate>Mon, 29 Sep 2025 19:45:01 -0500</pubDate>
       <guid>https://hasithv.github.io/posts/25-09-29-nanodiffgpt/</guid>
@@ -246,9 +246,10 @@ Hasting in a roseman.
 <p>It was a fun little experiment to convert nanoGPT into a diffusion LLM. I&rsquo;ve been thinking of other experiments on how to hijack architectures to make them do what I want, and so doing this was a good proof of concept as to how well it would work. Additionally, I got to explore diffusion LLMs, an area which I think holds tons of promise.</p>
 <h2 id="references">References</h2>
 <p>[1] GitHub repo - nanoDiffGPT: <a href="https://github.com/hasithv/nanoDiffGPT">https://github.com/hasithv/nanoDiffGPT</a></p>
-<p>[2] LLaDa paper: <a href="https://arxiv.org/abs/2502.09992">https://arxiv.org/abs/2502.09992</a></p>
-<p>[3] Block diffusion paper: <a href="https://arxiv.org/abs/2503.09573">https://arxiv.org/abs/2503.09573</a></p>
-<p>[4] MMaDa paper: <a href="https://arxiv.org/abs/2505.15809">https://arxiv.org/abs/2505.15809</a></p>
+<p>[2] Diffusion notes: <a href="/posts/flowdiffusion/flowdiff/">https://hasithv.github.io/posts/flowdiffusion/flowdiff/</a></p>
+<p>[3] LLaDa paper: <a href="https://arxiv.org/abs/2502.09992">https://arxiv.org/abs/2502.09992</a></p>
+<p>[4] Block diffusion paper: <a href="https://arxiv.org/abs/2503.09573">https://arxiv.org/abs/2503.09573</a></p>
+<p>[5] MMaDa paper: <a href="https://arxiv.org/abs/2505.15809">https://arxiv.org/abs/2505.15809</a></p>
 ]]></content:encoded>
     </item>
     <item>

public/index.xml

@@ -167,7 +167,7 @@ <h1>
 
 <article class="post-entry"> 
   <header class="entry-header">
-    <h2 class="entry-hint-parent">Making Nano-GPT a Diffusion LLM
+    <h2 class="entry-hint-parent">Hacking Nano-GPT into a Diffusion LLM
     </h2>
   </header>
   <div class="entry-content">
@@ -176,7 +176,7 @@ <h2 class="entry-hint-parent">Making Nano-GPT a Diffusion LLM
 ...</p>
   </div>
   <footer class="entry-footer"><span title='2025-09-29 19:45:01 -0500 CDT'>September 29, 2025</span>&nbsp;·&nbsp;10 min</footer>
-  <a class="entry-link" aria-label="post link to Making Nano-GPT a Diffusion LLM" href="https://hasithv.github.io/posts/25-09-29-nanodiffgpt/"></a>
+  <a class="entry-link" aria-label="post link to Hacking Nano-GPT into a Diffusion LLM" href="https://hasithv.github.io/posts/25-09-29-nanodiffgpt/"></a>
 </article>
 
 <article class="post-entry"> 

public/posts/25-09-29-nanodiffgpt/index.html

@@ -9,7 +9,7 @@
     <lastBuildDate>Mon, 29 Sep 2025 19:45:01 -0500</lastBuildDate>
     <atom:link href="https://hasithv.github.io/posts/index.xml" rel="self" type="application/rss+xml" />
     <item>
-      <title>Making Nano-GPT a Diffusion LLM</title>
+      <title>Hacking Nano-GPT into a Diffusion LLM</title>
       <link>https://hasithv.github.io/posts/25-09-29-nanodiffgpt/</link>
       <pubDate>Mon, 29 Sep 2025 19:45:01 -0500</pubDate>
       <guid>https://hasithv.github.io/posts/25-09-29-nanodiffgpt/</guid>
@@ -246,9 +246,10 @@ Hasting in a roseman.
 <p>It was a fun little experiment to convert nanoGPT into a diffusion LLM. I&rsquo;ve been thinking of other experiments on how to hijack architectures to make them do what I want, and so doing this was a good proof of concept as to how well it would work. Additionally, I got to explore diffusion LLMs, an area which I think holds tons of promise.</p>
 <h2 id="references">References</h2>
 <p>[1] GitHub repo - nanoDiffGPT: <a href="https://github.com/hasithv/nanoDiffGPT">https://github.com/hasithv/nanoDiffGPT</a></p>
-<p>[2] LLaDa paper: <a href="https://arxiv.org/abs/2502.09992">https://arxiv.org/abs/2502.09992</a></p>
-<p>[3] Block diffusion paper: <a href="https://arxiv.org/abs/2503.09573">https://arxiv.org/abs/2503.09573</a></p>
-<p>[4] MMaDa paper: <a href="https://arxiv.org/abs/2505.15809">https://arxiv.org/abs/2505.15809</a></p>
+<p>[2] Diffusion notes: <a href="/posts/flowdiffusion/flowdiff/">https://hasithv.github.io/posts/flowdiffusion/flowdiff/</a></p>
+<p>[3] LLaDa paper: <a href="https://arxiv.org/abs/2502.09992">https://arxiv.org/abs/2502.09992</a></p>
+<p>[4] Block diffusion paper: <a href="https://arxiv.org/abs/2503.09573">https://arxiv.org/abs/2503.09573</a></p>
+<p>[5] MMaDa paper: <a href="https://arxiv.org/abs/2505.15809">https://arxiv.org/abs/2505.15809</a></p>
 ]]></content:encoded>
     </item>
     <item>