{
  "schema_version": 1,
  "builder_plan": {
    "scenario": "Long customer conversations exhaust memory or limit concurrent requests.",
    "change": "Change only the attention cache policy, after checking model and runtime support.",
    "baseline": "The same model with its ordinary attention cache.",
    "steps": [
      "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.",
      "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.",
      "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.",
      "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test."
    ],
    "measure": "Peak cache memory, long-context answer quality, first-token latency, and requests completed.",
    "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation."
  },
  "builder_review": null,
  "arxiv_id": "2610.18026",
  "title": "ChampionKV: Head-Wise Adaptive Budget Allocation for KV Cache Compression",
  "summary": "ChampionKV replaces the uniform or predefined per-head budget allocation in observation-window KV compressors (SnapKV, PyramidKV) with input-adaptive head-wise budgets derived from attention concentration, leaving the token-selection procedure unchanged. The cost is negligible compute overhead for computing aggregated observation-window attention scores, plus an abnormal-head mask and floor-budget mechanism, with no additional memory beyond the compressed cache itself. It can fail to deliver gains on 7B/8B-class models, where results are mixed, and it does not uniformly outperform Ada-KV's global cross-head Top-K selection, so the base compressor choice may dominate the improvement.",
  "rationale": "It is a low-overhead plug-in the reader can run on a single 24 GB GPU, but the abstract reports statistically significant gains only on a 70B model and mixed results at the 7B/8B scale the reader is likely to deploy, so workload-specific validation is required before adoption.",
  "technique": "Head-wise adaptive KV cache budget allocation",
  "family": "cache_kv",
  "recommendation": "testar",
  "published": "2026-10-13",
  "as_of": "2026-10-18",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2610.18026",
  "independent_implementations": 0,
  "total_implementations": 0,
  "stars": 0,
  "citations": null,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "ChampionKV replaces the uniform or predefined per-head budget allocation in observation-window KV compressors (SnapKV, PyramidKV) with input-adaptive head-wise budgets derived from attention concentration, leaving the token-selection procedure unchanged. The cost is negligible compute overhead for computing aggregated observation-window attention scores, plus an abnormal-head mask and floor-budget mechanism, with no additional memory beyond the compressed cache itself. It can fail to deliver gains on 7B/8B-class models, where results are mixed, and it does not uniformly outperform Ada-KV's global cross-head Top-K selection, so the base compressor choice may dominate the improvement.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.18026",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "claim_id": "claim-02",
      "statement": "On LongBench and RULER, ChampionKV variants consistently improve over SnapKV and PyramidKV base compressors, with the largest and statistically significant gains on Llama3.1-70B-Instruct under tight budgets; results on 7B/8B models are competitive but mixed, and Ada-KV retains an advantage in several settings.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.18026",
      "source_page": null,
      "source_excerpt": ""
    }
  ],
  "exposure_map": [
    {
      "dimension": "quality",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "compute",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "latency",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "operations",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "compatibility",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "security",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "data_and_training",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "reproducibility",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    }
  ],
  "risks": [],
  "minimum_test": [],
  "open_questions": [],
  "independent_tests": [],
  "lab_reproductions": [],
  "report_available": false,
  "validation_tier": "unknown",
  "evidence_tier": "unknown",
  "infrastructure_basis": "unknown",
  "equations": [],
  "equations_status": "not_fetched",
  "core_kind": "",
  "equations_fetched_at": "",
  "equations_source_url": "",
  "equations_source_sha256": "",
  "exposures_reviewed_at": "",
  "exposures_source_sha256": ""
}