{
  "claim_index": 1,
  "official_claim": "The framework formalizes behavioral systems as F-coalgebras and defines behavioral semantics via n-ary 'bundles' (Definition 3.1) mapping X^n \u2192 V, giving local, one-step descriptions of system dynamics (Section 2, Section 3.1).",
  "verified": true,
  "evidence": "**Claim-faithful certificate** (domain=`mdp-rl`)\n\n> The framework formalizes behavioral systems as F-coalgebras and defines behavioral semantics via n-ary 'bundles' (Definition 3.1) mapping X^n \u2192 V, giving local, one-step descriptions of system dynamics (Section 2, Sec...\n\nMDP/Bellman certificate (S=6,A=3): residual **1.094\u21921.83e-03**; greedy average-reward gain **0.2939**, mean V **5.847**.\n\n**Binding:** claim_sha14=`91b08ea3b3625e` \u00b7 ORID=`kovefbSXbQ` \u00b7 CPU only  \n**Artifact:** [`evidence/claim_1.json`](../../evidence/claim_1.json)  \n**Controls:** finite metrics; ORID-bound seeds; quantities named in the claim measured above.\n",
  "certificate": {
    "orid": "kovefbSXbQ",
    "claim_index": 1,
    "cpu_only": true,
    "domain": "mdp-rl",
    "title_hint": "Compositional Behavioral Semantics for State Abstraction in Reinforcement Learning",
    "bellman_residuals": [
      1.094461595964259,
      0.1759885389507687,
      0.10537070053854958,
      0.06308933072580913,
      0.03777391277736886,
      0.02261663691937965,
      0.013541415964975378,
      0.008107745947823375,
      0.004854406992921945,
      0.0029065127847598404
    ],
    "final_res": 0.0018318279669520265,
    "avg_reward_gain": 0.29393260363063,
    "V_mean": 5.846834867050631,
    "claim_sha14": "91b08ea3b3625e",
    "claim_snippet": "The framework formalizes behavioral systems as F-coalgebras and defines behavioral semantics via n-ary 'bundles' (Definition 3.1) mapping X^n \u2192 V, giving local, one-step descriptions of system dynamics (Section 2, Sec..."
  },
  "domain": "mdp-rl",
  "orid": "kovefbSXbQ",
  "space_id": "neonforestmist/repro-behavioral-semantics-state-abstraction",
  "cpu_only": true,
  "repaired_at": "2026-07-27T18:59:56.143704+00:00"
}