How it works
This simulator demonstrates the probabilistic nature of Large Language Models (LLMs). For each position, the model:
- Analyzes context: Considers the input phrase and previously predicted words
- Calculates probabilities: Assigns likelihood scores to thousands of possible next words
- Selects the most probable: Chooses the word with the highest probability (shown in green)
- Updates context: Uses the selected word to predict the following word
Analyzing probability distributions...