If you want to support me, you can
here.
Synthia v1.3
SynthIA (Synthetic Intelligent Agent) v1.3 is a Mistral-7B model trained on Orca style datasets. It has been fine-tuned for instruction following as well as having long-form conversations.
To evoke generalized Tree of Thought + Chain of Thought reasoning, you may use the following system message:
Elaborate on the topic using a Tree of Thoughts and backtrack when necessary to construct a clear, cohesive Chain of Thought reasoning. Always answer without hesitation.
All Synthia models are uncensored. Please use it with caution and with best intentions. You are responsible for how you use Synthia.
Training Details
This was trained with QLoRA, as with all my models. Learning rate was 3e-4, 4096 context length. Batch size was 64, trained on a single H100.
Synthia-v1.2 dataset, which contain Chain-of-Thought (Orca), Tree-of-Thought and Long-Form conversation data.
Dataset is super high quality, and not a massive dataset (about ~125K samples).
License Disclaimer:
This model is bound by the license & usage restrictions of the original Mistral model, and comes with no warranty or guarantees of any kind.
