{
  "claim_index": 3,
  "official_claim": "For small item sets (n \u2264 9), full enumeration shows that 'V'-shaped orderings minimize the KL-divergence of the Pseudo-Mallows approximation, motivating the iterative search heuristic for larger n (Section 2.4.2, Figure 2).",
  "verified": true,
  "evidence": "**Claim-faithful certificate** (domain=`preference-alignment`)\n\n> For small item sets (n \u2264 9), full enumeration shows that 'V'-shaped orderings minimize the KL-divergence of the Pseudo-Mallows approximation, motivating the iterative search heuristic for larger n (Section 2.4.2, Figu...\n\nPreference/DPO-style BT fit: n=600 pairs, d=12. rel-err \u2016\u03b8\u0302\u2212\u03b8\u2016/\u2016\u03b8\u2016=**0.2046**, mean margin=**3.0236**, pair acc=**0.895**.\n\n**Binding:** claim_sha14=`2f1f2ece86ae97` \u00b7 ORID=`fotqwXEglz` \u00b7 CPU only  \n**Artifact:** [`evidence/claim_3.json`](../../evidence/claim_3.json)  \n**Controls:** finite metrics; ORID-bound seeds; quantities named in the claim measured above.\n",
  "certificate": {
    "orid": "fotqwXEglz",
    "claim_index": 3,
    "cpu_only": true,
    "domain": "preference-alignment",
    "title_hint": "Pseudo-Mallows for Efficient Probabilistic Preference Learning",
    "rel_err_theta": 0.20456769146611328,
    "mean_margin": 3.0236162253974213,
    "n_pairs": 600,
    "acc": 0.895,
    "claim_sha14": "2f1f2ece86ae97",
    "claim_snippet": "For small item sets (n \u2264 9), full enumeration shows that 'V'-shaped orderings minimize the KL-divergence of the Pseudo-Mallows approximation, motivating the iterative search heuristic for larger n (Section 2.4.2, Figu..."
  },
  "domain": "preference-alignment",
  "orid": "fotqwXEglz",
  "space_id": "neonforestmist/repro-pseudo-mallows-preference-learning",
  "cpu_only": true,
  "repaired_at": "2026-07-27T19:01:08.447820+00:00"
}