Generative AI fundamentals
Generative AI - AI systems that can create new content.
Traditional AI - AI systems that analyze and categorize existing data.
Examples:
- Traditional AI. It might classify emails as spam or not based on patterns.
- Generative AI. It can write a completely new email for you. LLMs are generative AI.
LLMs (Large Language Models) are trained to predict and generate human language. That's why they're called "language models". These models contain billions of parameters. That's why they're called "large". Parameters are mathematical values that determine how the model processes the information, somewhat like synaptic connections in your brain.
Three pillars that made the LLM possible
- Algorithms. There were algorithmic and architectural breakthroughs that fundamentally changed how AI systems learn. The development of transformer architecture in 2017 was a game changer.
- Data. There's a lot of digital data that AI models can learn on.
- Computation. Massive increase in computational power made it possible to train these complex models on all that data.
These factors led to an discovery called "scaling laws". As models grew larger and trained on more data with more computing power, their performance improved in predictable ways.
Steps of AI models training
Pre-training. You show the model text and ask it to predict what comes next. Through many iterations the model gradually refines its predictions. Thus it learns the patterns that make language coherent and meaningful.
Fine-tuning. They learn to follow instructions, provide helpful responses, avoid generating harmful content. This often involves human feedback to improve the model's performance. It also involves reinforcement learning which uses rewards and penalties to shape the model's behavior.
Once the models are trained, they are deployed so you can interact with them.
Important terms
Prompt. In order to interact with a model you provide a prompt. Prompt is a text that the model reads and then generates a response.
Context window. It's AI's working memory. It's a practical limit to how much information an LLM can consider at once. The context window includes your prompts, the AI responses, and any other information you've shared in your conversation.
Characteristics that make generative AI powerful
Ability to process vast amounts of information during training. That allows to learn complex patterns.
In-context learning ability. LLMs can adapt to new tasks based on instructions or examples in your prompt without requiring additional training.
Emerging capabilities that arise from scale. As the models grow larger, they develop capabilities that weren't explicitly designed into them.