Update README.md

README.md

@@ -11,10 +11,8 @@ AISAK, short for Artificially Intelligent Swiss Army Knife, is a state-of-the-ar
 
 - **Model Name**: AISAK
 - **Version**: 1.0
-- **Model Architecture**: Mixture of Experts (MoE)
-- **Specialization**: AISAK is structured upon the principles of the Mixture of Experts (MoE) architecture, meticulously crafted to emulate the success of the renowned https://huggingface.co/mistralai/Mixtral-8x7B-v0.1 model. Its architecture is ingeniously segmented into distinct expert modules, each adept at discerning specific patterns and features inherent within the input data.
-- **Gating Mechanism**:  A dynamic gating mechanism intelligently selects and combines the outputs of these experts based on the input data, enhancing adaptability and performance.
-- **Performance Comparison**:  While AISAK may not boast the same parameter count as the Mistral8x7b model, it maintains a remarkably high and heavily comparable performance level. Through meticulous optimization and leveraging the strengths of the MoE architecture, AISAK achieves results on par with its predecessor, ensuring that it stands as a formidable contender in the realm of artificial intelligence models.
+- **Model Architecture**: Transformer
+- **Specialization**: AISAK is structured upon the principles of the Transformer architecture, meticulously crafted to emulate the success of the renowned https://huggingface.co/mistralai/Mistral-7B-v0.1 model. Its architecture is ingeniously segmented into distinct expert modules, each adept at discerning specific patterns and features inherent within the input data.
 
 ### Intended Use:
 
@@ -22,7 +20,7 @@ AISAK, conceptualized by Mandela Logan, is intricately crafted for diverse text
 
 ### Performance:
 
-AISAK undergoes rigorous testing across diverse input data types, consistently demonstrating superior performance. Its capabilities have proven to outperform and exceed those of various state-of-the-art models such as but not limited to, GPT-3.5 and Llama 2 (70b).
+AISAK undergoes rigorous testing across diverse input data types, consistently demonstrating superior performance. Its capabilities have proven to outperform and exceed those of various state-of-the-art models such as but not limited to, GPT-3.5 and Llama 2's 13b and even 70b parameter model.
 
 ### Ethical Considerations: