{ "schema_version": 1, "builder_plan": { "scenario": "Your workload might fit a smaller or differently structured model.", "change": "Check for available weights and a compatible runtime before planning a product trial.", "baseline": "Your current model on identical held-out tasks and the same resource budget.", "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, latency, memory, context limits, and migration cost.", "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": "2609.40316", "title": "Scaling Laws for Looped Mixture of Experts", "summary": "The method replaces empirical, trial-and-error sizing of looped and MoE pretraining runs with a fitted scaling law that jointly predicts loss from recurrence depth, sparsity, model size, and data. Applying it requires pretraining models from scratch (validated at trillion-token scale), plus the added systems complexity of recurrence and expert routing; neither is available on a single 24 GB GPU. The fitted law can fail out of distribution: it is a predictive fit to a specific training regime, so extrapolation to other data mixes, objectives, or architectures may mispredict, and the claimed gains may not transfer to fine-tuned or API-based deployments.", "rationale": "The contribution is a pretraining design tool validated only at large scale, so it is relevant to follow but unusable without foundation-model training infrastructure.", "technique": "Joint scaling laws for looped MoE transformers", "family": "arquitetura_eficiente", "recommendation": "observar", "published": "2026-09-30", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.40316", "independent_implementations": 0, "total_implementations": 1, "stars": 0, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "The method replaces empirical, trial-and-error sizing of looped and MoE pretraining runs with a fitted scaling law that jointly predicts loss from recurrence depth, sparsity, model size, and data. Applying it requires pretraining models from scratch (validated at trillion-token scale), plus the added systems complexity of recurrence and expert routing; neither is available on a single 24 GB GPU. The fitted law can fail out of distribution: it is a predictive fit to a specific training regime, so extrapolation to other data mixes, objectives, or architectures may mispredict, and the claimed gains may not transfer to fine-tuned or API-based deployments.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.40316", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Sparsity delivers ~3x active-parameter efficiency; recurrence yields ~2x total-parameter efficiency on reasoning; at matched compute a looped MoE matches a ~2x larger non-looped MoE on reasoning benchmarks.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.40316", "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": "" }