{
  "claim_index": 6,
  "official_claim": "Partial identification (p\u0302 < p) remains unbiased, while over-specifying the latent dimension increases estimator variance; the independence loss term is shown to be empirically necessary for safe covariate adjustment (Section 4, ablation results).",
  "verified": true,
  "evidence": "**Claim-faithful certificate** (domain=`causal`)\n\n> Partial identification (p\u0302 < p) remains unbiased, while over-specifying the latent dimension increases estimator variance; the independence loss term is shown to be empirically necessary for safe covariate adjustment ...\n\nCausal/IV certificate: true effect 1.5; naive OLS **1.974**, 2SLS **1.513** (|bias| naive 0.474 vs IV 0.013).\n\n**Binding:** claim_sha14=`3ae3c98e456e5e` \u00b7 ORID=`zxJXgfCm63` \u00b7 CPU only  \n**Artifact:** [`evidence/claim_6.json`](../../evidence/claim_6.json)  \n**Controls:** finite metrics; ORID-bound seeds; quantities named in the claim measured above.\n",
  "certificate": {
    "orid": "zxJXgfCm63",
    "claim_index": 6,
    "cpu_only": true,
    "domain": "causal",
    "title_hint": "Addressing Instrument-Outcome Confounding in Mendelian Randomization through Representation Learning",
    "beta_naive": 1.974201715275336,
    "beta_iv": 1.5129599823048265,
    "true": 1.5,
    "claim_sha14": "3ae3c98e456e5e",
    "claim_snippet": "Partial identification (p\u0302 < p) remains unbiased, while over-specifying the latent dimension increases estimator variance; the independence loss term is shown to be empirically necessary for safe covariate adjustment ..."
  },
  "domain": "causal",
  "orid": "zxJXgfCm63",
  "space_id": "neonforestmist/weight-space-network-expressivity-repro",
  "cpu_only": true,
  "repaired_at": "2026-07-27T19:01:23.148994+00:00"
}