Paper Detail

It's Not RoPE that Creates Sinks: The Role of Self-Concentration and Value-Non-Mixing in Attention

Raito Kiya, Satoki Ohashi, Kosuke Sato, Go Kamoda, Ryosuke Takahashi, Yuji Yamamoto, Daiki Shiono, Keisuke Sakaguchi, Goro Kobayashi

arxiv Score 6.2

Published 2026-09-08 · First seen 2026-09-09

General AI

Abstract

Large Language Models (LLMs) often exhibit "Attention Sink" (AS) and the accompanying "Massive Activations" (MAs) at the initial position of a sequence. These phenomena frequently co-occur, and MAs can pose challenges for low-bit quantization. In this study, we analyze the factors underlying AS and MAs that emerge at the initial position regardless of the token occupying it. Our experiments suggest that self-concentration of attention, resulting from the causal mask, and the subsequent Value-non-mixing in attention outputs contribute to AS and MAs. These findings provide new empirical evidence on the internal dynamics of LLMs, offering insights that may inform future quantization strategies and advance our understanding of the internal mechanisms of attention layers.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
soon
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@article{kiya2026it,
  title = {It's Not RoPE that Creates Sinks: The Role of Self-Concentration and Value-Non-Mixing in Attention},
  author = {Raito Kiya and Satoki Ohashi and Kosuke Sato and Go Kamoda and Ryosuke Takahashi and Yuji Yamamoto and Daiki Shiono and Keisuke Sakaguchi and Goro Kobayashi},
  year = {2026},
  abstract = {Large Language Models (LLMs) often exhibit "Attention Sink" (AS) and the accompanying "Massive Activations" (MAs) at the initial position of a sequence. These phenomena frequently co-occur, and MAs can pose challenges for low-bit quantization. In this study, we analyze the factors underlying AS and MAs that emerge at the initial position regardless of the token occupying it. Our experiments suggest that self-concentration of attention, resulting from the causal mask, and the subsequent Value-non},
  url = {https://arxiv.org/abs/2609.09085},
  keywords = {cs.CL},
  eprint = {2609.09085},
  archiveprefix = {arXiv},
}

Metadata

{}