What is LLM?
LLM is a type of artificial intelligence algorithm that uses neural network techniques to understand human language or text using self-supervised learning techniques. LLM is purely based on deep learning Methodologies.
Deep Learning
Deep learning is changing how machines understand things, learn and interact with complex data. Deep learning uses neural networks that mimic the human brain to learn from complicated information.
Neural networks are machine learning models that mimic the complex functions of the human brain. These models consist of interconnected nodes or neurons that process data, learn patterns, and enable tasks such as pattern recognition and decision-making.
LLM Architecture
Large Language models, especially those using transformer architectures, have transformed the field by making text processing and generation highly accurate and relevant. The structure of these models is both advanced and intriguing, consisting of several essential components and mechanisms that enable their powerful performance.
LLM Layers
- Embedding layer
- Feedforward layer
- Recurrent layer
- Attention layer
Embedding Layer
It creates embedding from the input text. This part of the large language model captures the semantic and syntactic meaning of the input, so the model can understand context.
Feedforward Layer
Feedforward layer uses many connected layers to change the input word representations. This helps the model understand more abstract ideas.
Recurrent Layer
The recurrent layer processes the words in a sentence one by one, in the order they appear. It helps us understand how the words are connected and how they influence each other. This allows the model to recognize patterns and meaning within the sentence.
Attention Layer
The attention mechanism helps a language model concentrate on the most important parts of the input text based on the task it needs to complete. Instead of treating all words equally, this layer identifies and gives more importance to the words that matter most. By doing this, the model can understand the context better and produce more accurate results.
Use cases
We can perform text generation, summary writing, chatbot, image generation, conversational AI etc.
Examples of LLM models Chat GPT by open AI, BERT (Bidirectional Encoder Representations from Transformers) by Google, etc.
