EntmaxKV: Support-Aware Decoding for Entmax Attention

TL;DR

EntmaxKV introduces a support-aware sparse decoding method for entmax attention, significantly reducing memory traffic and increasing decoding speed in long-context language models while maintaining accuracy.

Contribution

It presents a novel entmax-native sparse decoding framework that exploits support recovery and adaptive candidate selection to improve efficiency over softmax-based methods.

Findings

Achieves up to 3.36x speedup over full attention at 1M context length.

Drops less probability mass and retains more support tokens than softmax-based methods.

Closely matches full-cache entmax performance with significantly reduced KV cache usage.

Abstract

Long-context decoding is increasingly limited by KV-cache memory traffic since each generated token attends over a cache whose size grows linearly with context length. Existing sparse decoding methods reduce this cost by selecting subsets of tokens or pages, but are designed for softmax attention, whose dense tails make any truncation discard nonzero probability mass. In contrast, $\alpha$-entmax produces exact zeros, turning sparse decoding from dense-tail approximation into support recovery: if the selected candidates contain the entmax support, sparse decoding remains exact. While recent entmax kernels enable efficient training, they do not address the autoregressive decoding bottleneck, where dense inference still streams the full KV cache before sparsity is known. In this work, we introduce EntmaxKV, an entmax-native sparse decoding framework that exploits sparsity before KV pages are loaded. EntmaxKV combines query-aware page scoring, support-aware candidate selection, and sparse entmax attention. We analyze truncation error through the dropped probability mass $\delta$, showing that output error is controlled by $\delta$ and vanishes when the entmax support is recovered. We further introduce a Gaussian-aware entmax selector that estimates the entmax threshold from lightweight page statistics, adapting the selected budget to the score distribution. Empirically, EntmaxKV drops less probability mass, retains more support tokens, and achieves lower output error than softmax-based sparse decoding at matched KV budgets. On long-context and language modeling benchmarks, it closely matches full-cache entmax while using a small fraction of the KV cache, achieving up to $3.36\times$ (softmax) and $5.43\times$ (entmax) speedup over full attention baselines at 1M context length. Code available at: https://github.com/deep-spin/entmaxkv.