{ "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": "2610.11790", "title": "Easy to anticipate, hard to compute: boundary dependence finds the computed outputs that entropy patching misses", "summary": "The method replaces BLT's entropy-triggered patch-start rule with a boundary-dependence signal (the rise in the model's own loss when a patch start is removed, measured per two-byte context), optionally combined with entropy, to place patch starts at positions whose values must be computed rather than merely predicted. It costs an extra loss-evaluation pass per candidate position to compute the dependence signal and still operates within the same patch budget, so global-model compute per byte is unchanged; gains are reported as accuracy points, not speed or memory. It fails when the underlying byte model cannot compute the target values at all, gives little benefit on copies and lookups, and only matters for workloads using byte-level patch-based models such as BLT rather than tokenized APIs.", "rationale": "The technique is a drop-in placement rule that requires only LoRA fine-tuning of a public 1B byte model, which fits a single 24 GB GPU, but it is worth adopting only after confirming the workload actually uses byte-level patching and contains computed-value positions, so it warrants a targeted small-scale test.", "technique": "Boundary-dependence patch placement for byte-level LMs", "family": "arquitetura_eficiente", "recommendation": "testar", "published": "2026-10-08", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.11790", "independent_implementations": 1, "total_implementations": 2, "stars": 122, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "The method replaces BLT's entropy-triggered patch-start rule with a boundary-dependence signal (the rise in the model's own loss when a patch start is removed, measured per two-byte context), optionally combined with entropy, to place patch starts at positions whose values must be computed rather than merely predicted. It costs an extra loss-evaluation pass per candidate position to compute the dependence signal and still operates within the same patch budget, so global-model compute per byte is unchanged; gains are reported as accuracy points, not speed or memory. It fails when the underlying byte model cannot compute the target values at all, gives little benefit on copies and lookups, and only matters for workloads using byte-level patch-based models such as BLT rather than tokenized APIs.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11790", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "On BLT-1B at a 10% patch budget, entropy placement gets 19.0% of GSM8K computed results exactly right, a boundary after each '=' gets 51.8%, and entropy plus label-free boundary dependence gets 67.1%; after LoRA adaptation, boundary dependence reaches 72.7% vs 32.9% for entropy (paired p < 1e-200).", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11790", "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": "" }