MHLA: Restoring Expressivity of Linear Attention via Token-Level Multi-Head

Kewei Zhang, Ye Huang, yfdeng, Jincheng Yu, Chen, Huan Ling, Enze Xie, Daquan Zhou

MHLA: Restoring Expressivity of Linear Attention via Token-Level Multi-Head: 40 upvotes on Hugging Face Daily Papers, #4 of 42 papers on 2026-01-13. Day-by-day upvote history. It lost 13 votes when the Hub removed votes in bulk.

While the Transformer architecture dominates many fields, its quadratic self-attention complexity hinders its use in large-scale applications. Linear attention offers an efficient alternative, but its direct application often degrades performance, with existing fixes typically re-introducing computational overhead through extra modules (e.g., depthwise separable convolution) that defeat the original purpose. In this work, we identify a key failure mode in these methods: global context collapse, where the model loses representational diversity. To address this, we propose Multi-Head Linear Attention (MHLA), which preserves this diversity by computing attention within divided heads along the token dimension. We prove that MHLA maintains linear complexity while recovering much of the expressive power of softmax attention, and verify its effectiveness across multiple domains, achieving a 3.6\% improvement on ImageNet classification, a 6.3\% gain on NLP, a 12.6\% improvement on image generation, and a 41\% enhancement on video generation under the same time complexity.

Paper page on Hugging Face · arXiv

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.