SSA: Sparse Sparse Attention by Aligning Full and Sparse Attention Outputs in Feature Space
By: Zhenyi Shen , Junru Lu , Lin Gui and more
Potential Business Impact:
Makes AI understand long stories better, faster.
The quadratic complexity of full attention limits efficient long-context processing in large language models (LLMs). Sparse attention mitigates this cost by restricting each query to attend to a subset of previous tokens; however, training-free approaches often lead to severe performance degradation. Native sparse-attention methods (e.g., NSA, MoBA) alleviate this issue, yet exhibit a critical paradox: they produce lower attention sparsity than full-attention models, despite aiming to approximate full attention, which may constrain their effectiveness. We attribute this paradox to gradient update deficiency: low-ranked key-value pairs excluded during sparse training receive neither forward contribution nor backward gradients, and thus never learn proper suppression. To overcome this limitation, we propose SSA (Sparse Sparse Attention), a unified training framework that considers both sparse and full attention and enforces bidirectional alignment at every layer. This design preserves gradient flow to all tokens while explicitly encouraging sparse-attention outputs to align with their full-attention counterparts, thereby promoting stronger sparsity. As a result, SSA achieves state-of-the-art performance under both sparse and full attention inference across multiple commonsense benchmarks. Furthermore, SSA enables models to adapt smoothly to varying sparsity budgets; performance improves consistently as more tokens are allowed to attend, supporting flexible compute-performance trade-offs at inference time. Finally, we show that native sparse-attention training surprisingly improves long-context extrapolation by mitigating the over-allocation of attention values in sink areas, with SSA demonstrating the strongest extrapolation capability.
Similar Papers
Optimizing Native Sparse Attention with Latent Attention and Local Global Alternating Strategies
Computation and Language
Makes computers understand long stories better.
Flash Sparse Attention: An Alternative Efficient Implementation of Native Sparse Attention Kernel
Distributed, Parallel, and Cluster Computing
Makes AI understand more words faster.
Making Every Head Count: Sparse Attention Without the Speed-Performance Trade-off
Machine Learning (CS)
Makes AI understand long texts much faster.