NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of ICLR 2027
163 results for “keyword:"sparse attention"” in ICLR 2027. 30 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]SudokuRAG: Efficient RAG Prefilling via Coordinated Intra- and Inter-Document Attention Sparsity▼4.3k2.00
[PDF]Triage Attention: Core-Protected Sparse Attention For Long-Video Vision-Language Models▼4.3k2.00
[PDF]VestigeKV: Query-Independent Sparse Attention with Certified Recall for NoPE-MLA Caches▼1611.50
[PDF]The Model Knows Where to Look: Distilling Attention Selectors for Long-Context Inference▼5.8k2.33
[PDF]Lost in a Few Heads: Diagnosing and Repairing Block Selection in Sparse Attention Language Models▲5103.00
[PDF]LazyFlash: Forward-Dense, Backward-Sparse Attention for Accelerated Long-Context Fine-Tuning▲4.7k3.25
[PDF]WASP: Weighted Asymmetric Sparse Attention to Accelerate both Prefilling and Decoding in Long-context Models▲3.0k2.33
[PDF]Oracle-Guided Sparse Prefill: Separating Oracle, Indexer, and Runtime Gaps in Long-Context GQA▲11.00
[PDF]HorizonKV: An Overlap-Aware Sparse Attention System for Long-Horizon Autoregressive Video Diffusion▼4.3k2.75
[PDF]BitLadder: Bit-Level Selective Loading of the KV Cache for Efficient Long-Context LLM Inference▼4.3k2.00
[PDF]RouteCache: Efficient Long-Context Attention via Learnable Retrieval and Temporal Caching▲3972.67
Withdrawn or desk rejected (30)
[PDF]BRAID: Band-Gate Routed Attention with Interleaved Detail for Efficient Video Diffusion TransformersDesk rejected
[PDF]Recovering Evicted Response for Progressive Sparse Attention in Image Super-ResolutionWithdrawn1.67
[PDF]SubCenSA: Training-Free Block-Sparse Decoding with Count-Weighted Sub-Centroid Mass EstimationWithdrawn1.67
[PDF]FLOYD VISION TRANSFORMER: RELATION-CENTRIC VISUAL BACKBONES WITH PIVOTAL ATTENTIONWithdrawn2.00
[PDF]LatentIndex: Cross-Layer Sharing with Layer-Specific Selection for Sparse AttentionWithdrawn2.00
[PDF]VASC: Value-Aware Sparse Attention with Cross-layer Memory for Efficient 3D ReconstructionWithdrawn1.67