TRANSFORMERS & LLMs • 52 / 397
Follow text from characters to tokens, attention, transformers and generation.

Decoder

A decoder generates output tokens, usually one token at a time based on previous context.

Think of it like

Think of Decoder as part of a very fast reader that repeatedly decides what information matters and what should come next.

Real life

Autocomplete, translation and chat assistants depend on concepts such as Decoder.

SRE lens

Chat-style LLMs are commonly decoder-oriented and produce streaming text.

Remember thisA decoder generates output tokens, usually one token at a time based on previous context.
AIForSREJump to a concept
Search is optional. The main journey is simply ↓ Next.