{
  "schema_version": 1,
  "builder_plan": {
    "scenario": "A self-hosted model uses more compute or memory than your workload can justify.",
    "change": "Check for a usable pruned checkpoint and runtime support before attempting model surgery.",
    "baseline": "The original unpruned model at the same precision, context, and load.",
    "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 quality, actual GPU memory, serving throughput, and preparation 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": "2609.06974",
  "title": "Train Overcomplete, Deploy Compact: Scaling Recovery Capacity for Structured LLM Pruning",
  "summary": "OverRep replaces standard compact recovery modules used after structured pruning with a temporarily overparameterized module that absorbs more distilled knowledge, then is algebraically merged back into a mathematically equivalent compact form via an annealed activation, leaving inference architecture and cost unchanged. The cost is additional training-time memory and compute for the overcomplete module plus implementation complexity of the reparameterization and merging procedure; inference-time memory and TFLOPs are unchanged. Failure modes include gains that may not transfer beyond the three evaluated backbone families or reasoning benchmarks, approximation error if the annealed activation does not fully converge to the linear regime, and no benefit unless the reader already performs structured pruning with a recovery-distillation stage.",
  "rationale": "The method only matters for teams already doing structured pruning with recovery training, and its quality gains are reported in points rather than a validated factor, so workload-specific small-scale validation is required before adoption.",
  "technique": "Overcomplete reparameterization for pruning recovery",
  "family": "esparsidade_e_poda",
  "recommendation": "testar",
  "published": "2026-09-07",
  "as_of": "2026-10-18",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2609.06974",
  "independent_implementations": 0,
  "total_implementations": 1,
  "stars": 2,
  "citations": 0,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "OverRep replaces standard compact recovery modules used after structured pruning with a temporarily overparameterized module that absorbs more distilled knowledge, then is algebraically merged back into a mathematically equivalent compact form via an annealed activation, leaving inference architecture and cost unchanged. The cost is additional training-time memory and compute for the overcomplete module plus implementation complexity of the reparameterization and merging procedure; inference-time memory and TFLOPs are unchanged. Failure modes include gains that may not transfer beyond the three evaluated backbone families or reasoning benchmarks, approximation error if the annealed activation does not fully converge to the linear regime, and no benefit unless the reader already performs structured pruning with a recovery-distillation stage.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2609.06974",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "claim_id": "claim-02",
      "statement": "Improves retained reasoning performance over strong recovery baselines by up to 5.5 and 8.4 points at 25% and 50% structured pruning, respectively, with memory and TFLOPs comparable to existing recovery methods.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2609.06974",
      "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_formula",
  "core_kind": "algorithm",
  "equations_fetched_at": "2026-09-27",
  "equations_source_url": "",
  "equations_source_sha256": "",
  "exposures_reviewed_at": "",
  "exposures_source_sha256": ""
}