{ "schema_version": 1, "builder_plan": { "scenario": "An agent looks good in demos, but you cannot tell which product tasks it fails.", "change": "Use the study's evaluation idea to inspect a small set of tasks with human-checked answers.", "baseline": "Your current evaluator and a manually reviewed reference set.", "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": "Agreement with human review, missed failures, false alarms, and evaluation cost per task.", "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": "2610.07782", "title": "Persistent Memory in Multi-Agent LLM Inference: What It Costs, What It Buys, and When You Can Tell", "summary": "A persistent memory tier storing and recalling reasoning traces replaces per-call re-derivation of intermediate reasoning, layered on top of multi-agent decomposition that bounds the active KV cache. It costs +0.368 MiB of peak cache per query plus the engineering complexity of trace storage, retrieval, and mandatory state reset between independent queries, with no measured accuracy benefit on single-question benchmarks. It fails when tasks supply their own evidence and are scored independently, since recalled traces then contain nothing informative, and poorly controlled ablations can hide this null behind measurement artifacts.", "rationale": "The evidence shows persistent memory adds cost without quality gains on standard single-question workloads, so it should only be adopted after workload-specific validation on genuinely stateful, cross-query tasks, while the KV-memory benefit of decomposition is directly relevant to a 24 GB GPU budget.", "technique": "Persistent memory tier in multi-agent inference", "family": "avaliacao_de_agente", "recommendation": "testar", "published": "2026-10-06", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.07782", "independent_implementations": 0, "total_implementations": 0, "stars": 0, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "A persistent memory tier storing and recalling reasoning traces replaces per-call re-derivation of intermediate reasoning, layered on top of multi-agent decomposition that bounds the active KV cache. It costs +0.368 MiB of peak cache per query plus the engineering complexity of trace storage, retrieval, and mandatory state reset between independent queries, with no measured accuracy benefit on single-question benchmarks. It fails when tasks supply their own evidence and are scored independently, since recalled traces then contain nothing informative, and poorly controlled ablations can hide this null behind measurement artifacts.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.07782", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Multi-agent decomposition reduces peak KV working set to 14.3 MiB per query versus 35.5 MiB (single-pass) and 35.3 MiB (retrieval-augmented) baselines, roughly 2.5x; the persistent memory tier adds +0.368 MiB with no detectable accuracy change (+0.015, 95% CI [-0.011, +0.046] across eight dataset pairs at n=100 per arm).", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.07782", "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": "" }