mlsanityData to 7 Oct 2026

CommitKV: Lifecycle-Aware KV Cache Compression via Commit Transitions for Multi-Turn Agents

arXiv 2608.07855 ↗First posted 8 Aug 2026cs.LGcited by 0

On the map: Efficient LLM inference and serving → Efficient reasoning and agent inference

Abstract (arXiv)

Multi-turn Reasoning-and-Acting (ReAct) agents accumulate growing trajectories of reasoning, tool calls, and observations. Their key-value (KV) caches grow accordingly, increasing memory use and attention cost during model inference. Existing KV cache compression methods reduce these costs by evicting states with low attention scores. However, low attention in the current turn does not imply future irrelevance, as temporarily inactive information may become important later. Snapshot-based eviction methods therefore do not explicitly distinguish temporarily dormant information from information that appears to have completed its role. In this paper, we present CommitKV, which identifies KV lifecycles through commit transitions. Specifically, CommitKV first divides completed agent events into token pages and compares each eligible page's deletion effect before a tool-call commit and after the commit's returned observation has been incorporated. Based on these paired measurements, CommitKV distinguishes dormant pages from high-to-low completion candidates. It then applies a greedy joint test, accepting candidates for retirement only when their combined post-commit effect remains bounded. Finally, at a later compression checkpoint, accepted pages are excluded, a bounded set of pages awaiting post-commit measurement is protected, and the remaining KV states are retained within the cache budget using the same token indices for keys, values, and absolute positions. These mechanisms ensure that CommitKV can distinguish dormant information from information that has completed its observed role and can be safely removed. Experiments on various benchmarks show that CommitKV reduces agent memory use, accelerates end-to-end inference, and achieves higher accuracy than existing KV cache compression methods.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Efficient reasoning and agent inference, one of 12 directions in Efficient LLM inference and serving. Builds on 59 of its 66 references that are in the map. First posted on arXiv 8 Aug 2026.

Builds on

59 of 66 references.

Closest ancestors · 6

Same direction, newest first

  1. 2026MemDecay: Region-Aware KV Cache Eviction for Efficient LLM Agent Inference12 Jul
  2. 2026IntentKV: Cross-Turn Intent-Aware KV Cache Pruning for Agent Inference6 Jun
  3. 2026Value-Aware Stochastic KV Cache Eviction for Reasoning Models2 Jun
  4. 2026TriAttention: Efficient Long Reasoning with Trigonometric KV Compression6 Apr
  5. 2026SideQuest: Model-Driven KV Cache Management for Long-Horizon Agentic Reasoning26 Feb
  6. 2025R-KV: Redundancy-aware KV Cache Compression for Reasoning Models30 May

Other references · 53

  1. 2017Attention Is All You Needcited by 59,932
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  3. 2025Qwen3 Technical Reportcited by 8,195
  4. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,146
  5. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  6. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,581
  7. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,748
  8. 2025gpt-oss-120b & gpt-oss-20b Model Cardcited by 1,258
  9. 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,107
  10. 2023GAIA: a benchmark for General AI Assistantscited by 1,084
  11. 2022Measuring and Narrowing the Compositionality Gap in Language Modelscited by 1,054
  12. 2024SnapKV: LLM Knows What You are Looking for Before Generationcited by 848
38 more
  1. 2023H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Modelscited by 805
  2. 2024Large Language Model based Multi-Agents: A Survey of Progress and Challengescited by 752
  3. 2024KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cachecited by 604
  4. 2024KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantizationcited by 571
  5. 2024Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inferencecited by 487
  6. 2023Model Tells You What to Discard: Adaptive KV Cache Compression for LLMscited by 466
  7. 2023Scissorhands: Exploiting the Persistence of Importance Hypothesis for LLM KV Cache Compression at Test Timecited by 428
  8. 2024PyramidKV: Dynamic KV Cache Compression based on Pyramidal Information Funnelingcited by 420
  9. 2024InternLM2 Technical Reportcited by 391
  10. 2024Ada-KV: Optimizing KV Cache Eviction by Adaptive Budget Allocation for Efficient LLM Inferencecited by 230
  11. 2024The Landscape of Emerging AI Agent Architectures for Reasoning, Planning, and Tool Calling: A Surveycited by 164
  12. 2024Cost-Efficient Large Language Model Serving for Multi-turn Conversations with CachedAttentioncited by 159
  13. 2024Fact, Fetch, and Reason: A Unified Evaluation of Retrieval-Augmented Generationcited by 158
  14. 2024GEAR: An Efficient KV Cache Compression Recipe for Near-Lossless Generative Inference of LLMcited by 155
  15. 2025Phi-4-reasoning Technical Reportcited by 146
  16. 2024ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inferencecited by 132
  17. 2024Exploring Large Language Model based Intelligent Agents: Definitions, Methods, and Prospectscited by 130
  18. 2024MiniCache: KV Cache Compression in Depth Dimension for Large Language Modelscited by 122
  19. 2025xbench: Tracking Agents Productivity Scaling with Profession-Aligned Real-World Evaluationscited by 118
  20. 2024Keep the Cost Down: A Review on Methods to Optimize LLM' s KV-Cache Consumptioncited by 96
  21. 2024PQCache: Product Quantization-based KVCache for Long Context LLM Inferencecited by 93
  22. 2024ZipCache: Accurate and Efficient KV Cache Quantization with Salient Token Identificationcited by 86
  23. 2024AffineQuant: Affine Transformation Quantization for Large Language Modelscited by 64
  24. 2024Palu: Compressing KV-Cache with Low-Rank Projectioncited by 59
  25. 2025KeyDiff: Key Similarity-Based KV Cache Eviction for Long-Context LLM Inference in Resource-Constrained Environmentscited by 49
  26. 2025ToolHop: A Query-Driven Benchmark for Evaluating Large Language Models in Multi-Hop Tool Usecited by 48
  27. 2021OMPQ: Orthogonal Mixed Precision Quantizationcited by 37
  28. 2026Flow caching for autoregressive video generationcited by 19
  29. 2021An Information Theory-inspired Strategy for Automatic Network Pruningcited by 17
  30. 2023Solving Oscillation Problem in Post-Training Quantization Through a Theoretical Perspectivecited by 17
  31. 2025Tool-R1: Sample-Efficient Reinforcement Learning for Agentic Tool Usecited by 16
  32. 2025Determining Layer-wise Sparsity for Large Language Models Through a Theoretical Perspectivecited by 13
  33. 2026SocialOmni: Benchmarking Audio-Visual Social Interactivity in Omni Modelscited by 12
  34. 2025Dynamic Low-Rank Sparse Adaptation for Large Language Modelscited by 10
  35. 2026HiMemVLN: Enhancing Reliability of Open-Source Zero-Shot Vision-and-Language Navigation with Hierarchical Memory Systemcited by 8
  36. 2026Self-GC: Self-Governing Context for Long-Horizon LLM Agentscited by 8
  37. 2025Rethinking Key-Value Cache Compression Techniques for Large Language Model Servingcited by 7
  38. 2025PagedEviction: Structured Block-wise KV Cache Pruning for Efficient Large Language Model Inferencecited by 7

50 most cited of 53 All 53 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026ActKV: Efficient LLM Agents through Action-Guided KV Cache Management
  2. 2026StepKV: Step-Aware KV Cache Compression for LLM Agents
  3. 2026Behavior-Preserving KV Cache Compression
  4. 2026Practical Online KV Cache Compaction for LLM Agents: An Empirical Study
  5. 2026AvoKV-E: Payload-Aware KV Cache Eviction for Long Reasoning
  6. 2026AgentKV: Phase-Aware KV Eviction for Agentic LLMs
  7. 2026KV-Fold: One-Step KV-Cache Recurrence for Long-Context Inference
  8. 2026AgentKVShift: Efficient KV Cache Reuse for Agentic Memory Systems