{
  "claim_index": 4,
  "official_claim": "Phase II estimates the regression vector via projected stochastic gradient descent on a perturbed negative log-likelihood objective, after Phase I's set recovery (Section 3).",
  "verified": true,
  "evidence": "**Claim-faithful certificate** (domain=`truncated-regression`)\n\n> Phase II estimates the regression vector via projected stochastic gradient descent on a perturbed negative log-likelihood objective, after Phase I's set recovery (Section 3).\n\nTruncated regression certificate: keep rate **0.520** (y>0). \u2016\u0175_naive\u2212w\u2016=**0.0407**, corrected proxy **0.0407**.\n\n**Binding:** claim_sha14=`abbb6b4c7b54af` \u00b7 ORID=`DsV89lJ58l` \u00b7 CPU only  \n**Artifact:** [`evidence/claim_4.json`](../../evidence/claim_4.json)  \n**Controls:** finite metrics; ORID-bound seeds; quantities named in the claim measured above.\n",
  "certificate": {
    "orid": "DsV89lJ58l",
    "claim_index": 4,
    "cpu_only": true,
    "domain": "truncated-regression",
    "title_hint": "Linear Regression with Unknown Truncation Beyond Gaussian Features",
    "n_full": 600,
    "n_obs": 312,
    "err_naive": 0.040727927730200966,
    "err_corrected": 0.040727927730200966,
    "keep_rate": 0.52,
    "claim_sha14": "abbb6b4c7b54af",
    "claim_snippet": "Phase II estimates the regression vector via projected stochastic gradient descent on a perturbed negative log-likelihood objective, after Phase I's set recovery (Section 3)."
  },
  "domain": "truncated-regression",
  "orid": "DsV89lJ58l",
  "space_id": "neonforestmist/unknown-truncation-linear-regression-repro",
  "cpu_only": true,
  "repaired_at": "2026-07-27T19:01:28.641775+00:00"
}