Large Language Model (LLM)
A large language model is a neural network trained on huge amounts of text to understand and generate human language. It predicts the next word (token) in a sequence, which lets it answer questions, write, and converse.
Data & AI
LLMs like the ones behind modern AI assistants are built on the transformer architecture and have billions of parameters. By learning statistical patterns from vast text, they can summarize, translate, write code, and hold conversations.
At their core, they're sophisticated next-token predictors: given the text so far, they estimate what word is likely to come next and repeat. Their abilities emerge from scale — more data, more parameters, and more compute.
Examples
Python
prompt = "The capital of France is"
# An LLM predicts the most likely next token
next_token = " Paris"
print(prompt + next_token)Output:
The capital of France is Paris