{ "schema_version": 1, "builder_plan": { "scenario": "Profiling shows attention computation is a bottleneck in your self-hosted service.", "change": "Check the supported GPU, tensor shapes, and precision, then replace only the relevant kernel.", "baseline": "The existing kernel inside the same complete serving pipeline.", "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": "Numerical differences, end-to-end latency, throughput, and memory at realistic shapes.", "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": "2610.30951", "title": "OpenAttention: One Percent Is Enough with Tail Proxies", "summary": "Replaces dense attention over the full KV cache with top-k key selection plus a count-aware mean proxy token that folds in the unread tail, using training-free query-geometry clustering, error-based read allocation (ErrorGate), and a custom flash kernel. Costs: an additional clustering/selection pass per decode step, a custom kernel dependency, and residual approximation error bounded by the proxy's fidelity to the true tail distribution. Can fail when the selector misses heavy-hitter keys whose mass the mean proxy poorly represents, on workloads with attention patterns unlike the long-context benchmarks used, or when deployment targets hardware the custom kernel does not support.", "rationale": "The method is training-free and targets exactly the reader's single-GPU inference setting, but the claimed 1% read accuracy and 10x kernel speedup require validation on the reader's models, context lengths, and hardware before production adoption.", "technique": "Sparse attention with count-aware mean tail proxy", "family": "kernels_e_atencao", "recommendation": "testar", "published": "2026-10-22", "as_of": "2026-10-23", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.30951", "independent_implementations": 0, "total_implementations": 1, "stars": 11, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "Replaces dense attention over the full KV cache with top-k key selection plus a count-aware mean proxy token that folds in the unread tail, using training-free query-geometry clustering, error-based read allocation (ErrorGate), and a custom flash kernel. Costs: an additional clustering/selection pass per decode step, a custom kernel dependency, and residual approximation error bounded by the proxy's fidelity to the true tail distribution. Can fail when the selector misses heavy-hitter keys whose mass the mean proxy poorly represents, on workloads with attention patterns unlike the long-context benchmarks used, or when deployment targets hardware the custom kernel does not support.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.30951", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Decode attention stage runs 10x faster than dense attention at a 1% read; retains 99% of dense accuracy on RULER-32K HARD with Llama-3.1-8B.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.30951", "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": "" }