Update README.md

Updates after model definition and tokenizer changes

README.md

@@ -22,7 +22,7 @@ license: mit
 - **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 73.32 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 73.32.
+It achieves an average score of 73.90 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 74.10.
 
 ### Model Optimizations
 
@@ -188,71 +188,71 @@ lm_eval \
   <tr>
    <td>MMLU (5-shot)
    </td>
-   <td>75.64
+   <td>76.69
    </td>
-   <td>76.04
+   <td>76.74
    </td>
-   <td>100.0%
+   <td>100.1%
    </td>
   </tr>
   <tr>
    <td>ARC Challenge (25-shot)
    </td>
-   <td>67.58
+   <td>69.45
    </td>
-   <td>69.71
+   <td>69.37
    </td>
-   <td>103.2%
+   <td>99.9%
    </td>
   </tr>
   <tr>
    <td>GSM-8K (5-shot, strict-match)
    </td>
-   <td>83.32
+   <td>85.22
    </td>
-   <td>82.03
+   <td>84.15
    </td>
-   <td>98.5%
+   <td>98.7%
    </td>
   </tr>
   <tr>
    <td>Hellaswag (10-shot)
    </td>
-   <td>84.37
+   <td>85.10
    </td>
-   <td>84.39
+   <td>84.76
    </td>
-   <td>100.0%
+   <td>99.6%
    </td>
   </tr>
   <tr>
    <td>Winogrande (5-shot)
    </td>
-   <td>75.45
+   <td>73.56
    </td>
-   <td>73.01
+   <td>73.80
    </td>
-   <td>96.8%
+   <td>100.3%
    </td>
   </tr>
   <tr>
    <td>TruthfulQA (0-shot)
    </td>
-   <td>53.54
+   <td>54.57
    </td>
-   <td>54.76
+   <td>54.57
    </td>
-   <td>102.3%
+   <td>100.0%
    </td>
   </tr>
   <tr>
    <td><strong>Average</strong>
    </td>
-   <td><strong>73.32</strong>
+   <td><strong>74.10</strong>
    </td>
-   <td><strong>73.32</strong>
+   <td><strong>73.90</strong>
    </td>
-   <td><strong>100.0%</strong>
+   <td><strong>99.7%</strong>
    </td>
   </tr>
 </table>