Discovering KV Cache Eviction Policies via LLM-Guided Program Evolution

[2608.14555] Discovering KV Cache Eviction Policies via LLM-Guided Program Evolution Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Computer Science > Distributed, Parallel, and Cluster Computing arXiv:2608.14555 (cs) [Submitted on 12 May 2026] Title: Discovering KV Cache Eviction Policies via LLM-Guided Program Evolution Authors: Pratik Poudel , Yanzhao Wu , Sumit Jha , Jason Liu View a PDF of the paper titled Discovering KV Cache Eviction Policies via LLM-Guided Program Evolution, by Pratik Poudel and 3 other authors View PDF HTML (experimental) Abstract: KV cache compression is critical for long-context inference, yet effective eviction policies remain difficult to design: existing prefill-stage methods often rely on hand-crafted salience heuristics that can be brittle across models, context lengths, and compression ratios. We present CacheCraft, a program-evolution methodology for automatically discovering KV cache eviction policies using an LLM-guided code-evolution engine. CacheCraft discovers FRC (Feature-Rich Compression), a fixed-weight three-signal scorer that combines local attention received, neighborhood attention density, and KV-head maximum salience with chunk-level top-k selection. Without per-model retuning, FRC ranks first among the evaluated single-pass KVPress baselines at every RULER 4k/8k cell with r >= 0.75 across Llama-3.1-8B-Instruct and Qwen3-8B (12 of 20 grid cells), gaining +15.4 points on Llama-4k and +13.9 points on Qwen-8k at 88% compression. A scorer-versus-structure decomposition shows that the scoring family, not chunk selection, is the load-bearing design choice: incorporating the scorer contributes +67.2 RULER points, while improving chunk structure contributes only ~0.1. Beyond FRC itself, CacheCraft provides a transferable recipe for automated eviction-policy discovery: a compact policy interface, a cascade evaluator with strict output invariants, and a diagnostic loop that treats search plateaus and reward-hacking failures as evidence for reformulating the editable interface. Subjects: Distributed, Parallel, and Cluster Computing (cs.DC) Cite as: arXiv:2608.14555 [cs.DC] (or arXiv:2608.14555v1 [cs.DC] for this version) https://doi.org/10.48550/arXiv.2608.14555 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Pratik Poudel [ view email ] [v1] Tue, 12 May 2026 20:23:20 UTC (6,804 KB) Full-text links: Access Paper: View a PDF of the paper titled Discovering KV Cache Eviction Policies via LLM-Guided Program Evolution, by Pratik Poudel and 3 other authors View PDF HTML (experimental) TeX Source view license Current browse context: cs.DC < prev | next > new | recent | 2026-08 Change to browse by: cs References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• Discovering KV Cache Eviction Policies via LLM-Guided Program Evolution#608996
• Aquifer: Hierarchical Memory Pooling with CXL and RDMA for MicroVM Snapshots#303184
• Debug Like a Human: Scaling LLM-based Fault Localization to Processor Design via Block-Level Instruction-Oriented Slicing#200570
• Master Ledger of Forensic Indebtedness: Sovereign Penalties for Unauthorized LLM Training and AI Data Extraction — Unearth Heritage Foundry#29021
• Master Ledger of Forensic Indebtedness: Sovereign Penalties for Unauthorized LLM Training and AI Data Extraction — Unearth Heritage Foundry#29010