{ "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": "2609.24698", "title": "Adapting Tree-Structured Speculative Decoding to DeepSeek-V4 for Efficient Inference", "summary": "Replaces linear (single-chain) speculative decoding with tree-structured speculation that retains multiple candidate branches per shared prefix, adapted to DeepSeek-V4's CSA/HCA compressed attention via branch-aware causal verification, temporary state isolation, and accepted-path state refresh. Costs additional draft/verify compute per step, more complex verification and KV-state management code, and requires integration into the DeepSeek-V4-Flash pipeline rather than an off-the-shelf serving stack. Gains are marginal at the smallest budget (D=5), throughput plateaus beyond a certain budget even as accepted length keeps rising, and benefits concentrate on less predictable workloads at small-to-medium batch sizes, so poorly matched workloads may see little improvement.", "rationale": "The speedup is real but modest and workload-dependent, and adoption requires the specific DeepSeek-V4-Flash pipeline plus nontrivial verification-state engineering, so the reader should validate on their own traffic before committing.", "technique": "Tree-structured speculative decoding for compressed-attention LLMs", "family": "decodificacao_especulativa", "recommendation": "testar", "published": "2026-09-21", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.24698", "independent_implementations": 0, "total_implementations": 0, "stars": 0, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "Replaces linear (single-chain) speculative decoding with tree-structured speculation that retains multiple candidate branches per shared prefix, adapted to DeepSeek-V4's CSA/HCA compressed attention via branch-aware causal verification, temporary state isolation, and accepted-path state refresh. Costs additional draft/verify compute per step, more complex verification and KV-state management code, and requires integration into the DeepSeek-V4-Flash pipeline rather than an off-the-shelf serving stack. Gains are marginal at the smallest budget (D=5), throughput plateaus beyond a certain budget even as accepted length keeps rising, and benefits concentrate on less predictable workloads at small-to-medium batch sizes, so poorly matched workloads may see little improvement.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.24698", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Up to about 18.5% throughput improvement over matched linear speculative decoding; accepted length rises from 2.39-2.84 (linear) to 2.83-3.41 (tree) at budget D=8 across GSM8K, MBPP, and ShareGPT.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.24698", "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": "" }