Update README.md

README.md

@@ -14,4 +14,7 @@ size_categories:
 
 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/63f2b7bcbe95ed4c9a9e7669/xYf0GGeL4sxo4NBA1WiGE.png)
 
-This is an example of a dataset that can be used to finetune a base llm to have dual thinking modes by adding flags to the system prompt. Here is an example of a chat template (https://huggingface.co/AIGym/smollm3-base-3b/blob/main/chat_template.jinja).
+This is an example of a dataset that can be used to finetune a base llm to have dual thinking modes by adding flags to the system prompt. Here is an example of a chat template (https://huggingface.co/AIGym/smollm3-base-3b/blob/main/chat_template.jinja).
+
+
+If you are interested in pretraining a base model checkout this example (https://github.com/vibe-coding-projects/build-nanogpt)