CommunityKV: Efficient Long-Context Decoding via Graph Partitioning
cs.LG
Submitted: 2026-09-30
Updated: 2026-09-30
Code: https://github.com/amazon-science/community-kv
Sources
- CodePlan: Repository-level Coding using LLMs and Planning
- Longformer: The Long-Document Transformer
- MagicPIG: LSH Sampling for Efficient LLM Generation
- Generating Long Sequences with Sparse Transformers
- FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning
- DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models
- LongNet: Scaling Transformers to 1,000,000,000 Tokens
- Plan-and-Act: Improving Planning of Agents for Long-Horizon Tasks
- Model Tells You What to Discard: Adaptive KV Cache Compression for LLMs
- The Llama 3 Herd of Models
- HyperAttention: Long-context Attention in Near-Linear Time
- Multipole Attention for Efficient Long Context Reasoning
- RULER: What's the Real Context Size of Your Long-Context Language Models?
- Reformer: The Efficient Transformer
- Maintaining Leiden Communities in Large Dynamic Graphs
- FreeKV: Boosting KV Cache Retrieval for Efficient LLM Inference
- MoBA: Mixture of Block Attention for Long-Context LLMs
- The Sparse Frontier: Sparse Attention Trade-offs in Transformer LLMs
- YaRN: Efficient Context Window Extension of Large Language Models
- LongCodeBench: Evaluating Coding LLMs at 1M Context Windows
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks