Update README.md

README.md

@@ -65,6 +65,12 @@ Nearly every base model that isn't finetuned for a specific task was trained on
 
 ```
 
+"Instruct" models have these special tokens:
+
+```
+<prompt> your prompt goes here <output> the model outputs a result here.
+```
+
 Some applications where I can imagine these being useful are: warm-starting very small encoder-decoder models, fitting a new scaling law that takes into account smaller models, or having a "fuzzy wrapper" around an API. They also could be usable on their own (for classification or other) when finetuned on more specific datasets. I don't expect the 3.3m models to be useful for any task whatsoever. Every model was trained on a singular GPU, either a RTX2060, RTX3060, or a T4. 
 
 I'd , uh , appreciate help in evaluating all these models probably with lm harness!!