{ "schema_version": 1, "builder_plan": { "scenario": "A support or data assistant repeats work across related customer requests.", "change": "Try the paper's memory change in a replay of past tasks, with separate memory per customer.", "baseline": "Your current memory pipeline plus a no-memory control using the same agent.", "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": "Task success, stale memories, retrieval misses, response latency, and all memory maintenance costs.", "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": "2610.00958", "title": "Role-aware Heuristic Episodic Attention for Conversational LLMs", "summary": "REA replaces naive full-history concatenation with a structured context policy: a persistent instruction prefix plus episodic memory that keeps user turns raw, compresses model replies, and heuristically selects, compresses, or omits each historical turn. It costs an extra retrieval-and-compression pipeline, prompt-assembly complexity, and some loss of verbatim assistant history, but adds no training and runs on a single 24 GB GPU or via third-party APIs. It can fail when heuristic turn selection drops information needed later, when reply compression loses details the user references, or when instruction identification misclassifies constraints, and the reported gains come from judge-scored benchmarks on small backbones that may not transfer to the reader's workload.", "rationale": "The method is inference-time prompt management requiring no training or cluster, but its heuristic retrieval and compression trade-offs require validation on the reader's own conversational workload before production adoption.", "technique": "Role-aware heuristic episodic context management", "family": "memoria_e_contexto", "recommendation": "testar", "published": "2026-10-01", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.00958", "independent_implementations": 1, "total_implementations": 2, "stars": 125, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "REA replaces naive full-history concatenation with a structured context policy: a persistent instruction prefix plus episodic memory that keeps user turns raw, compresses model replies, and heuristically selects, compresses, or omits each historical turn. It costs an extra retrieval-and-compression pipeline, prompt-assembly complexity, and some loss of verbatim assistant history, but adds no training and runs on a single 24 GB GPU or via third-party APIs. It can fail when heuristic turn selection drops information needed later, when reply compression loses details the user references, or when instruction identification misclassifies constraints, and the reported gains come from judge-scored benchmarks on small backbones that may not transfer to the reader's workload.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.00958", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "On Long-MT-Bench+, REA raises judge score from 6.32 to 7.36 (16.5% relative over the Vanilla baseline) and reduces average latency by 2.91x, with aggregate gains on 1.7B-7B backbones and Chinese/English role-playing tasks.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.00958", "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": "" }