Update README.md

README.md

@@ -19,7 +19,7 @@ pipeline_tag: text-generation
 - **Model Developers:** Neural Magic
 
 Quantized version of [Phi-3-medium-128k-instruct](https://huggingface.co/microsoft/Phi-3-medium-128k-instruct), a 14 billion-parameter open model trained using the Phi-3 datasets.
-It achieves an average score of 73.87 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 74.07.
+It achieves an average score of 73.99 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 73.32.
 
 ### Model Optimizations
 
@@ -181,71 +181,71 @@ lm_eval \
   <tr>
    <td>MMLU (5-shot)
    </td>
-   <td>76.61
+   <td>75.64
    </td>
-   <td>76.68
+   <td>76.79
    </td>
-   <td>100.1%
+   <td>101.5%
    </td>
   </tr>
   <tr>
    <td>ARC Challenge (25-shot)
    </td>
-   <td>69.37
+   <td>67.58
    </td>
-   <td>69.71
+   <td>69.54
    </td>
-   <td>100.5%
+   <td>102.9%
    </td>
   </tr>
   <tr>
    <td>GSM-8K (5-shot, strict-match)
    </td>
-   <td>84.31
+   <td>83.32
    </td>
-   <td>85.06
+   <td>84.31
    </td>
-   <td>100.0%
+   <td>101.2%
    </td>
   </tr>
   <tr>
    <td>Hellaswag (10-shot)
    </td>
-   <td>85.06
+   <td>84.37
    </td>
    <td>84.96
    </td>
-   <td>99.9%
+   <td>100.7%
    </td>
   </tr>
   <tr>
    <td>Winogrande (5-shot)
    </td>
-   <td>73.32
+   <td>75.45
    </td>
-   <td>73.32
+   <td>73.64
    </td>
-   <td>100.0%
+   <td>97.6%
    </td>
   </tr>
   <tr>
    <td>TruthfulQA (0-shot)
    </td>
-   <td>54.57
+   <td>53.54
    </td>
-   <td>54.68
+   <td>54.73
    </td>
-   <td>100.2%
+   <td>102.2%
    </td>
   </tr>
   <tr>
    <td><strong>Average</strong>
    </td>
-   <td><strong>73.87</strong>
+   <td><strong>73.32</strong>
    </td>
-   <td><strong>74.07</strong>
+   <td><strong>73.99</strong>
    </td>
-   <td><strong>100.3%</strong>
+   <td><strong>100.9%</strong>
    </td>
   </tr>
 </table>