Update README.md

README.md

@@ -24,6 +24,9 @@ use any gguf connector to interact with gguf file(s), i.e., [connector](https://
 - bug fixed following the guide written by [unsloth](https://unsloth.ai/blog/phi4)
 - tool used for quantization: [cutter](https://pypi.org/project/gguf-cutter)
 
+### citation
+[Phi-4 Technical Report](https://arxiv.org/pdf/2412.08905)
+
 ### appendices: model summary and quality (written by microsoft)
 
 #### model summary
@@ -58,6 +61,3 @@ to understand the capabilities, we (here refer to microsoft side) compare `phi-4
 | Reasoning                    | DROP          | 75.5      | 68.3            | 85.5                 | 79.3                 | **90.2**               | 76.7              | 80.9            |
 
 \* these scores are lower than those reported by Meta, perhaps because simple-evals has a strict formatting requirement that Llama models have particular trouble following.
-
-### citation
-[Phi-4 Technical Report](https://arxiv.org/pdf/2412.08905)