{ "schema_version": 1, "builder_plan": { "scenario": "You already need task-specific training and want to reduce its memory or compute cost.", "change": "Change only the training technique on a small controlled run after checking prerequisites.", "baseline": "Your current training recipe with the same model, data split, and evaluation.", "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": "Held-out quality, peak memory, training time, and full compute cost to the target quality.", "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": "2610.11160", "title": "LadderEdit: Edit-Level Residual Compression for Memory-Efficient Lifelong Editing of LLMs", "summary": "LadderEdit replaces storing one full-rank LoRA adapter per model edit, whose storage grows linearly with edit count, with a low-rank sketch per edit that is promoted up a rank ladder only if probe prompts fail the rewrite, generalization, and locality contract. The cost is a verification pass after each edit acquisition, extra rank for hard edits, and the engineering complexity of maintaining the probe suite and promotion logic; quality is claimed to track exact LoRA, not exceed it. Failure modes include probe prompts that under-cover real usage, letting contract-passing sketches degrade behavior in production, and memory savings eroding if the workload contains a high fraction of hard edits that get promoted to higher ranks.", "rationale": "The 5.2x memory reduction is relevant to a 24 GB single-GPU setup only if the reader actually performs continual model editing, and the claimed quality parity depends on probe coverage that must be validated on their own edit distribution.", "technique": "Rank-ladder compression of per-edit LoRA adapters", "family": "treino_eficiente", "recommendation": "testar", "published": "2026-10-08", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.11160", "independent_implementations": 0, "total_implementations": 0, "stars": 0, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "LadderEdit replaces storing one full-rank LoRA adapter per model edit, whose storage grows linearly with edit count, with a low-rank sketch per edit that is promoted up a rank ladder only if probe prompts fail the rewrite, generalization, and locality contract. The cost is a verification pass after each edit acquisition, extra rank for hard edits, and the engineering complexity of maintaining the probe suite and promotion logic; quality is claimed to track exact LoRA, not exceed it. Failure modes include probe prompts that under-cover real usage, letting contract-passing sketches degrade behavior in production, and memory savings eroding if the workload contains a high fraction of hard edits that get promoted to higher ranks.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11160", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Matches exact per-edit LoRA storage quality at 5.2x less memory and remains effective at 50,000 sequential edits on ZsRE, CounterFact, and WikiBigEdit with LLaMA-3-8B, Mistral-7B, and Qwen2.5-7B.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11160", "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": "" }