Updated on 2024-12-30

Author: lxaw

index.html

@@ -35,7 +35,7 @@ <h1>Where?</h1>
             <!-- Last Update Section -->
             <section>
                 <h1>When?</h1>
-        Last time this was edited was 2024-12-29 (YYYY/MM/DD).
+        Last time this was edited was 2024-12-30 (YYYY/MM/DD).
             </section>
 
             <!-- Footer -->

papers/list.json

@@ -1,4 +1,13 @@
 [
+  {
+    "title": "Turning Up the Heat: Min-p Sampling for Creative and Coherent LLM Outputs",
+    "author": "Nguyen Nhat Minh et al",
+    "year": "2024",
+    "topic": "sampling, llm",
+    "venue": "Arxiv",
+    "description": "This paper introduces a neat trick to sample the next token. Min-p sampling basically adjusts the sampling threshold based on the model's confidence. It does so by scaling according to the top token's probability. This is a compelling alternative to other common sampling methods, like nucleus sampling.",
+    "link": "https://arxiv.org/pdf/2407.01082"
+  },
   {
     "title": "LASER: Attention with Exponential Transformation",
     "author": "Sai Surya Duvvuri et al",

papers_read.html

@@ -16,10 +16,10 @@ <h1>Here's where I keep a list of papers I have read.</h1>
         I typically use this to organize papers I found interesting. Please feel free to do whatever you want with it. Note that this is not every single paper I have ever read, just a collection of ones that I remember to put down.
     </p>
     <p id="paperCount">
-        So far, we have read 196 papers. Let's keep it up!
+        So far, we have read 197 papers. Let's keep it up!
     </p> 
     <small id="searchCount">
-        Your search returned 196 papers. Nice! 
+        Your search returned 197 papers. Nice! 
     </small>
 
     <div class="search-inputs">
@@ -46,6 +46,16 @@ <h1>Here's where I keep a list of papers I have read.</h1>
         </thead>
         <tbody>
 
+            <tr>
+                <td>Turning Up the Heat: Min-p Sampling for Creative and Coherent LLM Outputs</td>
+                <td>Nguyen Nhat Minh et al</td>
+                <td>2024</td>
+                <td>sampling, llm</td>
+                <td>Arxiv</td>
+                <td>This paper introduces a neat trick to sample the next token. Min-p sampling basically adjusts the sampling threshold based on the model&#x27;s confidence. It does so by scaling according to the top token&#x27;s probability. This is a compelling alternative to other common sampling methods, like nucleus sampling.</td>
+                <td><a href="https://arxiv.org/pdf/2407.01082" target="_blank">Link</a></td>
+            </tr>
+        
             <tr>
                 <td>LASER: Attention with Exponential Transformation</td>
                 <td>Sai Surya Duvvuri et al</td>