Update README.md

README.md

@@ -42,6 +42,9 @@ Please note that the model was trained on unfiltered internet data. Since we do
 
 Bonus: The model underwent some fine-tuning on the prompt format introduced in LLaVA1.5 that is unrelated to image attention calculation. Therefore, aligning the ViT Projection module with frozen LMunder visual instructions would enable rapid implementation of effective multimodal capabilities.
 
+## PROMPT FORMAT:
+[chatml](https://github.com/openai/openai-python/blob/main/chatml.md)
+
 ## MMLU:
 stem ACC: 56.83 
 
@@ -54,17 +57,17 @@ social ACC: 72.41
 **AVERAGE ACC:63.82**
 
 ## CEval (Val):
-STEM ACC: 66.71 
+STEM acc: 61.67
 
-Social Science ACC: 85.10 
+Social Science acc: 81.94
 
-Humanities ACC: 76.68 
+Humanities acc: 77.19
 
-Other ACC: 70.23 
+Other acc: 68.35
 
-Hard ACC:54.71 
+Hard acc:48.03
 
-**AVERAGE ACC:73.10**
+**AVERAGE acc:70.27**
 
 ## GSM8K