{
  "schema_version": 1,
  "builder_plan": {
    "scenario": "You already need task-specific training and want to reduce its memory or compute cost.",
    "change": "Change only the training technique on a small controlled run after checking prerequisites.",
    "baseline": "Your current training recipe with the same model, data split, and evaluation.",
    "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": "Held-out quality, peak memory, training time, and full compute cost to the target quality.",
    "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": "2609.19754",
  "title": "AutoData: Agentic Search for Pre-training Data Selection",
  "summary": "AutoData replaces hand-designed pre-training data curation heuristics (mixture weights, quality filters) with an LLM agent that iteratively searches a program space of scoring, stratification, and stochastic selection rules, refined against validation feedback from a small proxy model. The cost is a search loop requiring repeated proxy-model pre-training runs (an 'overnight search' in the paper), plus agent orchestration and evaluation infrastructure, which exceeds a single-GPU constrained budget. The method can fail through overfitting the selection recipe to the proxy model or proxy benchmark, non-transfer to the reader's data distribution, and selection rules that exploit quirks of per-document features (lexical statistics, perplexity) rather than true data quality.",
  "rationale": "The technique targets pre-training data selection and requires repeated proxy-model training runs, which the reader's constrained infrastructure cannot support, though the agentic-data-engineering framing is worth monitoring.",
  "technique": "Agentic search over pre-training data selection algorithms",
  "family": "treino_eficiente",
  "recommendation": "observar",
  "published": "2026-09-17",
  "as_of": "2026-10-18",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2609.19754",
  "independent_implementations": 0,
  "total_implementations": 1,
  "stars": 8,
  "citations": null,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "AutoData replaces hand-designed pre-training data curation heuristics (mixture weights, quality filters) with an LLM agent that iteratively searches a program space of scoring, stratification, and stochastic selection rules, refined against validation feedback from a small proxy model. The cost is a search loop requiring repeated proxy-model pre-training runs (an 'overnight search' in the paper), plus agent orchestration and evaluation infrastructure, which exceeds a single-GPU constrained budget. The method can fail through overfitting the selection recipe to the proxy model or proxy benchmark, non-transfer to the reader's data distribution, and selection rules that exploit quirks of per-document features (lexical statistics, perplexity) rather than true data quality.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2609.19754",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "claim_id": "claim-02",
      "statement": "The discovered selection algorithm outperforms human-designed curation pipelines and improves the downstream CORE metric, with transfer from a small proxy model to larger scales; no multiplicative factor is reported.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2609.19754",
      "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": ""
}