Update README.md

README.md

@@ -13,4 +13,18 @@ This model has 10 layers, 10 heads and 640 embeddings, with a context window of
 It was able to achieve a training loss of 2.3256 and validation loss of 2.3651.
 Supervised fine-tuning should be performed before use.
 
-Note: GPT-Usenet uses MBOX syntax. For finetuning, your data should be in the .mbox format.
+Note: GPT-Usenet uses MBOX syntax. 
+```
+uucp:
+Internet:
+Path:
+From:
+Sender:
+Newsgroups:
+Subject:
+Message-ID:
+Date:
+Organization:
+Lines:
+```
+For finetuning, your data should be in the .mbox format.