<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Verification on Tan Ke</title>
    <link>https://mrtanke.github.io/tags/verification/</link>
    <description>Recent content in Verification on Tan Ke</description>
    <generator>Hugo -- 0.161.1</generator>
    <language>en-us</language>
    <lastBuildDate>Fri, 01 May 2026 10:11:25 +0000</lastBuildDate>
    <atom:link href="https://mrtanke.github.io/tags/verification/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Math-Shepherd</title>
      <link>https://mrtanke.github.io/posts/2026-05-01-math-shepherd/</link>
      <pubDate>Fri, 01 May 2026 10:11:25 +0000</pubDate>
      <guid>https://mrtanke.github.io/posts/2026-05-01-math-shepherd/</guid>
      <description>&lt;p&gt;This is the talk and presentation I’ve given during seminar &amp;ldquo;Process Reward Modeling in LLMs&amp;rdquo; at the University of Heidelberg.&lt;/p&gt;
&lt;p&gt;It involves a presentation and a short academic discussions, the content is about paper sharing, experiments, and reproduction results with classmates and professors.&lt;/p&gt;
&lt;p&gt;The paper name is &amp;ldquo;Math-Shepherd: Verify and Reinforce LLMs Step-by-Step without Human Annotations&amp;rdquo; (&lt;a href=&#34;https://arxiv.org/abs/2312.08935&#34; target=&#34;_blank&#34; rel=&#34;noopener noreferrer&#34;&gt;Wang et al., 2024&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;You could find the paper and slide here:&lt;/p&gt;
&lt;p&gt;&lt;a href=&#34;Math-Shepherd_paper.pdf&#34;&gt;Annotated Paper (PDF)&lt;/a&gt; |
&lt;a href=&#34;Math-Shepherd_ppt.pdf&#34;&gt;Preview Slides (PDF)&lt;/a&gt;&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
