AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Towards Understanding Massive Activations in Attention Sink Mechanism

Haiyu Wang, Yuanyuan Lin

ICML 2026 · Architectures and efficiency › Attention and transformers · approach: Sparse and structured attention

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
zero out massive activations of the sink token (ablation)pretrained LLM (LLaMA3.1-8B) · sink rate eps=0.3Sink_1 (%): 98.91 -> 9.19Hurt
additionally zero the sink token value vector (h = v = 0) after zeroing massive activations (ablation)pretrained LLM (LLaMA3.1-8B) · sink rate eps=0.3Sink_1 (%): 9.19 -> 96.60Improved
additionally zero the sink value vector after zeroing massive activations (ablation)LLaMA2-7B · WikiTextperplexity: 9774.18 -> 29.04 (original 5.37)Improved
zero out top-k massive activations before RMSNorm (ablation)LLaMA2-7B, LLaMA3-8B · sink rateSink_1 (%): decreases sharply to near zeroHurt
zero out top-k largest entries after RMSNorm (ablation)LLaMA2-7B, LLaMA3-8B · sink rateSink_1 (%): moderate dropHurt
rescale the sink hidden state in all subsequent RMSNorm layers (All layers) (ablation)LLaMA2-7B, LLaMA3-8B · sink rateSink_1 (%)No clear effect
rescale the sink hidden state at only one layer (One layer) (ablation)LLaMA2-7B, LLaMA3-8B · sink rateSink_1 (%)Hurt
remove massive activations from sink tokens (ablation)Llama2-7B · token perturbation propagationrepresentational distanceHurt
increase sink token hidden-state norm (ablation)LLaMA2-7B, LLaMA3-8B · Jacobian norm of sink tokenJacobian normImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy