Short Answer

Uniqueness of RoPE-based Embeddings

A language model generates a final, position-aware embedding, ei\mathbf{e}_i, by applying a rotational transformation to a token's initial embedding, xi\mathbf{x}_i, based on its position, ii. The process is described by the function ei=Ro(xi,iθ)\mathbf{e}_i = \mathrm{Ro}(\mathbf{x}_i, i\theta). If two different tokens (with distinct initial embeddings xA\mathbf{x}_A and xB\mathbf{x}_B) are located at the same position pp, is it possible for them to have identical final embeddings (i.e., eA=eB\mathbf{e}_A = \mathbf{e}_B)? Explain your reasoning based on the properties of a rotational transformation.

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

Ch.3 Prompting - Foundations of Large Language Models

Analysis in Bloom's Taxonomy

Cognitive Psychology

Psychology

Social Science

Empirical Science

Science