Update README.md

README.md

@@ -28,7 +28,7 @@ license: llama3.1
 - **Model Developers:** Neural Magic
 
 Quantized version of [Meta-Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-405B-Instruct).
-It achieves an average score of 86.01 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 86.63.
+It achieves an average score of 86.47 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 86.63.
 
 ### Model Optimizations
 
@@ -148,9 +148,9 @@ This version of the lm-evaluation-harness includes versions of ARC-Challenge, GS
    </td>
    <td>86.25
    </td>
-   <td>85.97
+   <td>86.17
    </td>
-   <td>99.67%
+   <td>99.90%
    </td>
   </tr>
   <tr>
@@ -158,9 +158,9 @@ This version of the lm-evaluation-harness includes versions of ARC-Challenge, GS
    </td>
    <td>96.93
    </td>
-   <td>95.39
+   <td>95.3
    </td>
-   <td>98.41%
+   <td>98.31%
    </td>
   </tr>
   <tr>
@@ -168,9 +168,9 @@ This version of the lm-evaluation-harness includes versions of ARC-Challenge, GS
    </td>
    <td>96.44
    </td>
-   <td>95.83
+   <td>96.05
    </td>
-   <td>99.36%
+   <td>99.59%
    </td>
   </tr>
   <tr>
@@ -178,9 +178,9 @@ This version of the lm-evaluation-harness includes versions of ARC-Challenge, GS
    </td> 
    <td>88.33
    </td>
-  <td>88.16
+   <td>88.27
    </td>
-   <td>99.80%
+   <td>99.93%
    </td>
   </tr>
   <tr>
@@ -188,9 +188,9 @@ This version of the lm-evaluation-harness includes versions of ARC-Challenge, GS
    </td>
    <td>87.21
    </td>
-  <td>85.95
+  <td>87.76
    </td>
-   <td>98.55%
+   <td>100.63%
    </td>
   </tr>
   <tr>
@@ -198,9 +198,9 @@ This version of the lm-evaluation-harness includes versions of ARC-Challenge, GS
    </td>
    <td>64.64
    </td>
-   <td>64.75
+   <td>65.27
    </td>
-   <td>100.17%
+   <td>100.97%
    </td>
   </tr>
   <tr>
@@ -208,9 +208,9 @@ This version of the lm-evaluation-harness includes versions of ARC-Challenge, GS
    </td>
    <td><strong>86.63</strong>
    </td>
-   <td><strong>86.01</strong>
+   <td><strong>86.47</strong>
    </td>
-   <td><strong>99.28%</strong>
+   <td><strong>99.81%</strong>
    </td>
   </tr>
 </table>