Fill in the Blank

An attention mechanism groups relative token distances (offsets) into buckets using the following rules:

  1. Offsets 0 through 15 are mapped directly to their corresponding buckets (e.g., offset 12 is in bucket 12).
  2. For larger distances, the buckets cover logarithmically increasing ranges:
    • Bucket 16 covers offsets 16-20.
    • Bucket 17 covers offsets 21-26.
    • Bucket 18 covers offsets 27-33.

Following this pattern, the relative position offset of 30 would be assigned to bucket ____.

0

1

Updated 2025-10-04

Contributors are:

Who are from:

Tags

Ch.2 Generative Models - Foundations of Large Language Models

Foundations of Large Language Models

Foundations of Large Language Models Course

Computing Sciences

Application in Bloom's Taxonomy

Cognitive Psychology

Psychology

Social Science

Empirical Science

Science

Related