Update README.md

README.md

@@ -1,13 +1,10 @@
 ---
 license: apache-2.0
-pipeline_tag: text-generation
-language:
-  - en
-tags:
-- pretrained
-inference:
-  parameters:
-    temperature: 0.7
 ---
+The prompt format is Vicuna/Sharegpt. 
 
-I'm sorry I don't have a real README yet. This model is trained on a subset of orca-best and heavily cleaned (but extremely degenerate) reverse proxy data logs. It uses Vicuna prompt format and is based off of Mistral-7b. I don't want to waste anyone's time until I'm positive I have something that can blow everyone's mind.
+This model was trained on a subset of orca-best combined with most of freedom-rp. To put it as a ratio, the dataset is roughly 90% orca-best and about 10% degenerate reverse proxy logs. The goal was to create a model with the intellegence and capacity of orca-best, but with enhanced roleplay and adult content capabilities. If you are looking for a model that is trained purely on wanton degeneracy without any attempt to retain intellegence, check out my cockatrice model.
+
+You can find the dataset here used to train this model here:
+
+https://huggingface.co/datasets/openerotica/basilisk-v0.2