{ "schema_version": 1, "builder_plan": { "scenario": "Your self-hosted model spends too long generating answers token by token.", "change": "Check whether a compatible draft model and serving integration exist before renting compute.", "baseline": "The same target with speculation off, then your existing drafter if you have one.", "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": "Accepted draft length, tokens per second, tail latency, output quality, and total GPU 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": "2610.23613", "title": "Trajectory-Triggered Sparse Collaborative Decoding", "summary": "Replaces dense speculative decoding's per-token target verification with a sparse scheme where the draft generates most tokens and a lightweight logistic predictor over aggregate trajectory-shape confidence features decides when to invoke the target for short corrective segments. Costs: training the trigger predictor on draft correctness labels, maintaining a draft-target pair, and added orchestration complexity; target compute drops substantially, but absolute latency savings versus dense baselines are not quantified in the abstract. Fails when local early evidence is too weak to predict trajectory failure (the stated timing tradeoff) or when the target model lacks competence on the task, in which case the method loses to open-loop baselines, as observed on MATH-500.", "rationale": "The accuracy gains are workload- and pair-dependent, and the method requires training a trigger and validating draft-target competence alignment on the reader's specific tasks before production adoption.", "technique": "Trajectory-triggered sparse collaborative decoding", "family": "decodificacao_especulativa", "recommendation": "testar", "published": "2026-10-16", "as_of": "2026-10-19", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.23613", "independent_implementations": 0, "total_implementations": 1, "stars": 122, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "Replaces dense speculative decoding's per-token target verification with a sparse scheme where the draft generates most tokens and a lightweight logistic predictor over aggregate trajectory-shape confidence features decides when to invoke the target for short corrective segments. Costs: training the trigger predictor on draft correctness labels, maintaining a draft-target pair, and added orchestration complexity; target compute drops substantially, but absolute latency savings versus dense baselines are not quantified in the abstract. Fails when local early evidence is too weak to predict trajectory failure (the stated timing tradeoff) or when the target model lacks competence on the task, in which case the method loses to open-loop baselines, as observed on MATH-500.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.23613", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Learned sparse triggering improves over the best open-loop schedule by +5.3 to +20.0 accuracy points across five GSM8K draft-target pairs, +7.0 to +13.5 points on MATH-500 when trigger and target align, and +5.0 to +8.3 points on StrategyQA.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.23613", "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": "" }