@inproceedings{wang-etal-2026-logical,
    title = "From Logical to Computational Sparsity: Structure-Aware Block-Sparse Attention for Long-Code Completion",
    author = "Wang, Yanli  and
      Wang, Yanlin  and
      Zhang, Bowen  and
      Zhang, Yiwei  and
      Guo, Daya  and
      Chen, Jiachi  and
      Zhang, Hongyu  and
      Zheng, Zibin",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.1189/",
    pages = "25927--25942",
    ISBN = "979-8-89176-390-6",
    abstract = "Code Large Language Models face critical Time-To-First-Token (TTFT) latency challenges when handling long code completion due to the quadratic complexity ($O(n^2)$) of attention mechanisms. While existing sparse attention methods attempt to address this issue, they suffer from three key limitations: (1) general sparse patterns cause excessive accuracy degradation without considering code structure, (2) code-specific methods achieve only logical sparsity without actual computational speedup, and (3) limited adaptation to complex scenarios such as repository-level completion. We propose **SabreCoder**, a training-free **S**tructure-**a**ware **b**lock-spa**r**s**e** attention mechanism that bridges the gap between logical and computational sparsity. SabreCoder parses code into semantic chunks, constructs chunk-level sparse patterns through dependency analysis and similarity matching, and maps them to GPU-friendly block-sparse formats. Extensive experiments on LCC and CrossCodeEval benchmarks demonstrate that SabreCoder reduces TTFT by 45-55{\%} while maintaining accuracy within 3{\%} of dense attention."
}