Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
140 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

The Rotary Position Embedding May Cause Dimension Inefficiency in Attention Heads for Long-Distance Retrieval (2502.11276v1)

Published 16 Feb 2025 in cs.CL and cs.LG

Abstract: The Rotary Position Embedding (RoPE) is widely used in the attention heads of many LLMs (LLM). It rotates dimensions in the query and the key vectors by different angles according to their positions in the input sequence. For long context modeling, the range of positions may vary a lot, and thus RoPE rotates some dimensions by a great range of angles. We hypothesize that the wide range of rotation angles may prevent LLMs from utilizing those dimensions. To validate this hypothesis, we present a controlled experiment showing that applying RoPE causes low utility of certain dimensions. Our analyses on three LLMs also indicate that these dimensions do not help LLMs do long-context question answering.

Summary

We haven't generated a summary for this paper yet.