{ "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": "2610.11784", "title": "Compile the Table: Query-Calibrated Operator Compression for Tabular In-Context Learning", "summary": "QCOC replaces storing raw in-context examples (or per-query retrieval subsets) in the KV cache of tabular ICL models with a one-time-compiled set of joint-KV prototypes shared across queries. The cost is an offline compilation pass with clustering and closed-form value calibration, plus implementation complexity beyond standard inference stacks, and a small residual accuracy gap (about 0.23 points). It can fail if the workload has few repeated queries over the same table (compilation cost is not amortized), if the underlying tabular ICL model or attention structure differs from the evaluated setup, or if cluster fidelity degrades on tables unlike the OpenML benchmarks.", "rationale": "The method offers a real throughput gain for teams serving tabular ICL models (e.g., TabPFN-style) at query volume, but it only applies to that niche workload and its accuracy retention must be validated on the reader's own tables before production adoption.", "technique": "Query-calibrated KV cache compression for tabular ICL", "family": "cache_kv", "recommendation": "testar", "published": "2026-10-08", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.11784", "independent_implementations": 0, "total_implementations": 1, "stars": 4, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "QCOC replaces storing raw in-context examples (or per-query retrieval subsets) in the KV cache of tabular ICL models with a one-time-compiled set of joint-KV prototypes shared across queries. The cost is an offline compilation pass with clustering and closed-form value calibration, plus implementation complexity beyond standard inference stacks, and a small residual accuracy gap (about 0.23 points). It can fail if the workload has few repeated queries over the same table (compilation cost is not amortized), if the underlying tabular ICL model or attention structure differs from the evaluated setup, or if cluster fidelity degrades on tables unlike the OpenML benchmarks.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11784", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Compressing 8,192 in-context examples to 512 memory slots gives a 10.5x cache compression ratio; online serving is up to 508x faster than dynamic retrieval baselines and 1.98x faster than full-context inference, with accuracy on average 0.23 percentage points below full context.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11784", "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": "" }