adding readme

README.md

@@ -4,9 +4,21 @@ license: apache-2.0
 
 [<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
 
+# Synthia-MoE-v3-Mixtral-8x7B
 
+This is Synthia-MoE-v3 trained on the official Mistral MoE version (Mixtral-8x7B). 
 
-This is Synthia trained on the official Mistral MoE version (Mixtral-8x7B).
+This model is trained on the Synthia-v3.0 dataset, that contains ~10K super high-quality GPT-4-Turbo generated samples. The samples contains Tree-of-Thought, Chain-of-Thought and other system contexts designed to evoke reasoning, philosophical thinking, use working memory and long chain of reasoning with multi-part questions.
+
+Further, this model is trained on the Orca-2 principle of replacing the system context with just one message. In the case of this Synthia-MoE-v3 model, the system context was not included at all.
+
+The evals are coming, but testing empirically the model produces highly intelligent, coherent results. Here's a sample conversation: https://migel.substack.com/p/a-conversation-with-synthia-moe-mixtral
+
+<br>
+
+![Synthia](https://huggingface.co/migtissera/Synthia-MoE-v3-Mixtral-8x7B/resolve/main/Synthia-MoE.png)
+
+<br>
 
 ```
 import torch, json