Learn Before
Formula

Augmented Input Formula for External Memories

When utilizing external memories or datastores for language modeling, the retrieved information is integrated with the original input to form an augmented input for the Large Language Model (LLM). If the original input is denoted as x\mathbf{x} and the retrieved context elements are c1,...,ck\mathbf{c}_{1}, ..., \mathbf{c}_{k}, they are combined using a function or template gg. The resulting augmented input, x\mathbf{x}', is mathematically expressed as: x=g(c1,...,ck,x)\mathbf{x}' = g(\mathbf{c}_{1}, ..., \mathbf{c}_{k}, \mathbf{x}). The accompanying illustration demonstrates how these external memories are integrated into the language modeling workflow.

Image 0

0

1

Updated 2026-04-23

Contributors are:

Who are from:

Tags

Foundations of Large Language Models

Ch.2 Generative Models - Foundations of Large Language Models

Foundations of Large Language Models Course

Computing Sciences