{ "schema_version": 1, "builder_plan": { "scenario": "An assistant chooses the wrong API or sends arguments your product cannot use.", "change": "Apply the proposed tool-selection or argument-generation change in a sandbox.", "baseline": "Your current agent with the same tool schemas and task 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": "Correct completed actions, invalid calls, retries, latency, and cost per completed task.", "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": "2610.21412", "title": "JevAgentHarness: Benchmarking, Deploying and Training Decision Models That Share Control with LLMs", "summary": "The harness replaces the large LLM call made at every multiple-choice step (e.g., tool selection) with a 4B decision model, falling back to the LLM when a per-choice confidence threshold fitted on the agent's own trajectories is not met. It requires running a 4B model alongside the agent (feasible on a 24 GB GPU), collecting trajectories to fit thresholds, and ideally fine-tuning the decision model, adding calibration and integration complexity. It can fail because isolated-question accuracy does not transfer to in-loop control (the best model solved only 69.5% of tasks, and uncalibrated delegation cut success by 19.0 points), and gains shrink on tasks unlike the decision model's training data.", "rationale": "The approach targets exactly the reader's constraint of reducing per-call LLM cost, but the measured 19-point success drop without per-choice threshold calibration means adoption is only safe after validating the harness on the reader's own workloads.", "technique": "Small decision model replaces per-step LLM calls", "family": "uso_de_ferramenta", "recommendation": "testar", "published": "2026-10-15", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.21412", "independent_implementations": 4, "total_implementations": 5, "stars": 1637, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "The harness replaces the large LLM call made at every multiple-choice step (e.g., tool selection) with a 4B decision model, falling back to the LLM when a per-choice confidence threshold fitted on the agent's own trajectories is not met. It requires running a 4B model alongside the agent (feasible on a 24 GB GPU), collecting trajectories to fit thresholds, and ideally fine-tuning the decision model, adding calibration and integration complexity. It can fail because isolated-question accuracy does not transfer to in-loop control (the best model solved only 69.5% of tasks, and uncalibrated delegation cut success by 19.0 points), and gains shrink on tasks unlike the decision model's training data.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.21412", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "On in-distribution tasks, 4B decision models made agents 81.1-89.0% cheaper and up to 2.5x faster, with task success loss under 3 points; out-of-distribution success was 91.5% with the trained model versus 82.0% with the base decision model.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.21412", "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": "" }