{
  "schema_version": 1,
  "builder_plan": {
    "scenario": "A multi-step product task fails because the agent loses track of dependencies.",
    "change": "Try the proposed planning step on tasks with checkable intermediate outcomes.",
    "baseline": "The same agent solving the same tasks without the new planning step.",
    "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 completion, unnecessary steps, planning tokens, response time, and total 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": "2610.12061",
  "title": "When Should Agents Think? Adaptive Reasoning via Cross-Turn Estimation",
  "summary": "RACE replaces the default policy of generating a reasoning step before every agent action with a learned policy that reasons only when prior reasoning no longer supports the next action, using likelihood drops on reference actions as a cheap removal signal. The cost is a training pipeline: the LoGiC detection procedure requires likelihood evaluation over reference trajectories, followed by supervised fine-tuning and agentic reinforcement learning, plus reference action data that may not exist for custom tasks. It can fail when the likelihood-drop signal misestimates cross-turn support, causing the policy to skip reasoning on turns where new reasoning is actually needed, and the gains may not transfer to domains or base models different from the training setup.",
  "rationale": "The approach directly targets inference cost on constrained infrastructure, but adopting it requires SFT plus agentic RL and reference-action data, so it warrants a small-scale validation on the reader's own workloads rather than immediate production adoption.",
  "technique": "Adaptive per-turn reasoning gating via likelihood estimation (RACE)",
  "family": "planejamento_e_decomposicao",
  "recommendation": "testar",
  "published": "2026-10-08",
  "as_of": "2026-10-18",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2610.12061",
  "independent_implementations": 1,
  "total_implementations": 2,
  "stars": 132,
  "citations": null,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "RACE replaces the default policy of generating a reasoning step before every agent action with a learned policy that reasons only when prior reasoning no longer supports the next action, using likelihood drops on reference actions as a cheap removal signal. The cost is a training pipeline: the LoGiC detection procedure requires likelihood evaluation over reference trajectories, followed by supervised fine-tuning and agentic reinforcement learning, plus reference action data that may not exist for custom tasks. It can fail when the likelihood-drop signal misestimates cross-turn support, causing the policy to skip reasoning on turns where new reasoning is actually needed, and the gains may not transfer to domains or base models different from the training setup.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.12061",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "claim_id": "claim-02",
      "statement": "RACE substantially reduces reasoning cost while maintaining or improving task performance on four agent benchmarks; no specific numeric factor is given in the abstract.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.12061",
      "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": ""
}