Update README.md

README.md

@@ -14,9 +14,9 @@ athirdpath/LD-Zephyria-37b_HEALED, continually pretrained on one epoch of:
 - Private Iambe dataset (1.6 GB)
 
 
-He named himself
+This model has NOT had instruction tuning after pretraining, use -INSTRUCT instead for chat/instruct. This model will do so (see below, Mistral format) but is best for creative writing.
 
-I've just taken to letting them do that 0-shot on deterministic unless the name is taken.
+He named himself - I've just taken to letting them do that 0-shot on deterministic unless the name is taken.
 
 <p align="center"><img src="https://cdn-uploads.huggingface.co/production/uploads/633a809fa4a8f33508dce32c/l7UxCJsOkqbZdcujx5s14.png"/></p>