Short Answer

Input Vector Creation in Autoregressive Generation

An autoregressive language model is generating text. After producing a token at a specific position in the sequence, the model must create a single vector to represent that token before feeding it into its processing layers. Describe the two distinct types of information that are typically combined to form this initial input vector.

0

1

Updated 2025-10-07

Contributors are:

Who are from:

Tags

Ch.5 Inference - Foundations of Large Language Models

Foundations of Large Language Models

Foundations of Large Language Models Course

Computing Sciences

Comprehension in Revised Bloom's Taxonomy

Cognitive Psychology

Psychology

Social Science

Empirical Science

Science