Research map / Architectures and efficiency / Attention and transformers
Max-Margin Token Selection in Attention Mechanism
Davoud Ataee Tarzanagh, Yingcong Li, Xuechen Zhang, Samet Oymak
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use normalized gradient descent (eta=0.1) instead of constant-step gradient descent (eta=1) | Softmax attention model (synthetic) · Synthetic token selection | time to reach softmax probability 1 | Improved |
| train longer so attention weight norm grows | ViT-base on CIFAR-10 · CIFAR-10 CLS attention map | attention sparsity | Improved |
| correlation loss instead of logistic loss | Attention model transient dynamics · Synthetic token scores | token bias | Mixed |
Closest papers by meaning
- Attention with Trained Embeddings Provably Selects Important Tokens (NeurIPS 2025 · similarity 0.90)
- Sparse Attention as Compact Kernel Regression (NeurIPS 2026 · similarity 0.88)
- From Attention to Activation: Unraveling the Enigmas of Large Language Models (ICLR 2025 · similarity 0.88)
- SAS: Simple Attention Sparsification via End-to-End Optimization of Context Ranking (NeurIPS 2026 · similarity 0.88)
- Benign Overfitting in Token Selection of Attention Mechanism (ICML 2025 · similarity 0.88)
- Beyond Linear Approximations: A Novel Pruning Approach for Attention Matrix (ICLR 2025 · similarity 0.87)
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.
