Update README.md

README.md

@@ -100,9 +100,9 @@ The dataset is comprised of a mixture of open datasets large-scale datasets avai
 ## Other benchmark:
 | Metric                | Value                     |
 |-----------------------|---------------------------|
-| ARC (25-shot)         |  0.47       |
-| HellaSwag (10-shot)   | 0.7421    |
-| MMLU (5-shot)        |   0.4627     |
+| ARC (25-shot)         |  0.470       |
+| HellaSwag (10-shot)   | 0.742    |
+| MMLU (5-shot)        |   0.463     |
 | TruthfulQA (0-shot)   |    |
 | Winogrande (5-shot)   |   0.655 |
 | GSM8K (5-shot)        | 0.423        |