Techniques & Methods
Generation
Generation in AI refers to the output phase: a trained model produces novel content conditioned on an input. For language models, generation is autoregressive—tokens are produced one at a time, with each token conditioned on the prompt and all previously generated tokens.
Controlling generation quality involves temperature (randomness), top-p and top-k sampling (vocabulary truncation), and length constraints. High temperature produces diverse, creative outputs; low temperature produces focused, deterministic ones.
Authority Links
Related Terms
Techniques & Methods
Sequence Generation
Process where models produce sequences—such as words or tokens—based on learned patterns.
Techniques & Methods
Autoregression
Statistical modeling approach where future values are predicted from past observed values.
Core Concepts
Natural Language Generation (NLG)
Generating coherent, contextually relevant text from structured data or prompts.
Techniques & Methods
Decoding Rules
Guidelines and algorithms that control how language models translate internal representations into output tokens.