Update README.md

README.md

@@ -199,10 +199,11 @@ Best results were reached with `contains_book` evaluation type. See the figure a
 | GPT  | cv | 0 |
 
 </details>
+<br>
 
 # Code
 
-Access to models via OpenRouter
+## Access to models via OpenRouter
 
 ```python
 import requests
@@ -230,7 +231,7 @@ def evaluate_openrouter(prompt: str, model_name: str) -> str:
   return data["choices"][0]["message"]["content"]
 ```
 
-Check correct
+## Check correct
 
 ```python
 def check_correct_contains(answer: str, llm_answer: str) -> bool: