Exact Fusion of Softmax Attention Heads Can Require Exponentially Many Atoms
cs.LG, cs.NE, stat.ML
Submitted: 2026-11-02
Updated: 2026-11-02
Code: https://github.com/usamihiroyasu/attention-head-fusion
Terminology
Sources
- The risk of KV cache compression
- Fast Transformer Decoding: One Write-Head is All You Need
- Attention-based representations for multi-task computation
- The Effect of Attention Head Count on Transformer Approximation
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis