Update README.md

README.md

@@ -131,20 +131,43 @@ pinned: true
         <li>Innovative data preparation and curation techniques</li>
         <li>Focus on domain-specific excellence and versatility</li>
         <li>Open collaboration and knowledge sharing within the AI community</li>
+        <li>Advancing LLM research via novel techniques, which has been applied by all of our members.</li>
       </ul>
       <p></p>
       <h2>Team:</h2>
       <div class="team-member">
+        <p>Chief Engineer</p>
         <p><strong>@elinas</strong> - <a href="https://huggingface.co/elinas" target="_blank">HuggingFace Profile</a></p>
       </div>
       <div class="team-member">
+        <p>Data Scientist</p>
+        <p><strong>@ToastyPigeon</strong> - <a href="https://huggingface.co/ToastyPigeon" target="_blank">HuggingFace Profile</a></p>
+      </div>
+      <div class="team-member">
+        <p>Ops Engineer</p>
+        <p><strong>@fizz</strong> - <a href="https://huggingface.co/Fizzarolli" target="_blank">HuggingFace Profile</a></p>
+      </div>
+      <div class="team-member">
+        <p>ML / DS Engineer</p>
         <p><strong>@SteelSkull</strong> - <a href="https://huggingface.co/Steelskull" target="_blank">HuggingFace Profile</a></p>
       </div>
       <p></p>
       <h2>Notable Achievements</h2>
       <div class="achievements">
         <ul>
-          <li>Development of L3-Aethora-15B series, The first heavily fintuned 15b model that focuses in creative writing and general intelligence.</li>
+          <li>Revival of Llama 1 33B by training on over 500M tokens</li>
+          <li>We did this based on the original pretraining token count of 1.4T and decided to add another 500M tokens to it, to which our surprise ended up
+            surpassing expectations in both quality and length</li>
+          <li>
+            It was trained at 16384 context legth with an *effective* context legnth around 12k due to the nature of the samples, but exceeds in RP.
+          </li>
+          <li>
+            Our next goal is to apply GQA to it, but in the meantime, we will appreciate quanters who will help with running this model on less VRAM!
+          </li>
+        </ul>
+        <ul>
+          <li>Development of L3-Aethora-15B series, The first heavily fintuned 15b model that focuses in creative writing and general intelligence using a novel
+            technique known as "zeroing layers."</li>
           <li>Creation of the Aether-Lite-V1.8.1 dataset, a carefully curated dataset for AI training</li>
         </ul>
       </div>