·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Predictive Self-Supervised Learning Provably Identifies Stochastic Signals under Nuisance1h◆Boosting Adversarial Robustness and Generalization with Dictionary Structure1h◆A Dominant Supplier Slows Recursive Drift More Than It Steers It1h◆COMiT: Learning Structured Visual Tokens through Sequential Communication1h◆Compress to Remember: Learning Compact Memory via On-Policy Distillation for Long Video Generation1h◆Data-Free Pruning of Self-Attention Layers in LLMs1h◆Screening Is Enough1h◆Data Unlearning via Inverse Distillation1h◆ReCIRC: Rectified Conformal Risk Control1h◆Replay-buffer engineering for noise-aware quantum circuit optimization1h◆SINO: Scale-Invariant Neural Operator1h◆Preferent Compression Bounds Are Tight1h◆Fundamental Limits of Transferability and Equivariance in Algebraic Signal Models I: Finite Dimensions1h◆How to Tame a Multi-Headed Hydra? Adaptive Multi-Category Safety Steering for Large Language Models1h◆Bregman Consensus1h◆In-Context Learning for Robots: Methods and Applications1h◆Modal Logic Neural Networks1h◆Volatility-Clustering Adaptation for Financial Time Series1h◆Reference-Guided Machine Unlearning1h◆Agentic Federated Learning: Rule-Based Client and Server Agents for Adaptive Training1h◆Predictive Self-Supervised Learning Provably Identifies Stochastic Signals under Nuisance1h◆Boosting Adversarial Robustness and Generalization with Dictionary Structure1h◆A Dominant Supplier Slows Recursive Drift More Than It Steers It1h◆COMiT: Learning Structured Visual Tokens through Sequential Communication1h◆Compress to Remember: Learning Compact Memory via On-Policy Distillation for Long Video Generation1h◆Data-Free Pruning of Self-Attention Layers in LLMs1h◆Screening Is Enough1h◆Data Unlearning via Inverse Distillation1h◆ReCIRC: Rectified Conformal Risk Control1h◆Replay-buffer engineering for noise-aware quantum circuit optimization1h◆SINO: Scale-Invariant Neural Operator1h◆Preferent Compression Bounds Are Tight1h◆Fundamental Limits of Transferability and Equivariance in Algebraic Signal Models I: Finite Dimensions1h◆How to Tame a Multi-Headed Hydra? Adaptive Multi-Category Safety Steering for Large Language Models1h◆Bregman Consensus1h◆In-Context Learning for Robots: Methods and Applications1h◆Modal Logic Neural Networks1h◆Volatility-Clustering Adaptation for Financial Time Series1h◆Reference-Guided Machine Unlearning1h◆Agentic Federated Learning: Rule-Based Client and Server Agents for Adaptive Training1h◆
News/Data-Free Pruning of Self-Attention Layers in LLMs
arxiv
PublishedOctober 1, 2026 at 4:00 AM
—neutral

Data-Free Pruning of Self-Attention Layers in LLMs

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2512.20636v2 Announce Type: replace Abstract: Many self-attention sublayers in large language models (LLMs) can be removed with little to no loss. We attribute this to the Attention Suppression Hypothesis: during pre-training, some deep attention layers learn to mute their own contribution, le

Stay posted· Newsletter

A 5-min weekly brief — top movers, price watch, story of the week.

// no spam · unsubscribe one-click · free forever

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →

Related coverage

More from ARXIV
arxivPredictive Self-Supervised Learning Provably Identifies Stochastic Signals under Nuisance1harxivBoosting Adversarial Robustness and Generalization with Dictionary Structure1harxivA Dominant Supplier Slows Recursive Drift More Than It Steers It1harxivCOMiT: Learning Structured Visual Tokens through Sequential Communication1h
The Bubble Brief
WEEKLY

Read AI insights every Tuesday — top movers, new releases, story of the week.

// no spam · unsubscribe one-click · free forever

Originally published on arxiv ↗
Built by Marouane Gazouzi
HomeModelsNews