{
 "page": "P148",
 "title": "Directional assimilation in accelerator anomaly classifiers: the ACRB hypotheses",
 "description": "The Accelerator Classifier Retention Battery sets out seven standing hypotheses on directional assimilation in learned anomaly selection: reconstruction-specific failure, cross-family persistence, the normalized-autoencoder remedy, the low-complexity blind spot, blind-spot inheritance through distillation, drift under quantization and firmware, and the value of architectural plurality. Inversion Battery v0.2 reports that only 36 to 47 percent of a teacher's top one percent survives into the student's, with the loss in distillation, that architecture changes the orientation of the blind spot, and what is deferred to v0.3. The priors paper and the OAR protocol name the program's contribution and its pre-registered class-conditional inversion battery at matched background acceptance. The common experiment proposed is to calibrate each model at matched operating points and measure how much of a withheld class is assimilated as ordinary.",
 "line": "accelerator-metrology",
 "url": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses",
 "claims": [
  {
   "id": "PRED-773.G1",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-05-31",
   "registered_in": 773,
   "registered_title": "The Cut Between Two Measures: On the Quantization Seam Joining the Directionality of Semantic Labor to the Deviation Family (EA-KURO-BRIDGE-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/773/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-773.G1",
   "statement": "DSL is the taxonomy-quantized, intent-baselined member of the deviation family. The statement is exact on-axis and divergent off-axis, and the divergence is not noise but the signature of the quantization — a testable prediction about *where* the two measures will separate on any case, not only this one.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F773.G1",
     "deposit": 773,
     "date": "2026-05-31",
     "lines": [
      53,
      53
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/773/",
     "quote": "DSL is the taxonomy-quantized, intent-baselined member of the deviation family. The statement is exact on-axis and divergent off-axis, and the divergence is not noise but the signature of the quantization — a testable prediction about *where* the two measures will separate on any case, not only this one.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-931.F1",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-06-29",
   "registered_in": 931,
   "registered_title": "EA-SEI-OAR-PROTOCOL v0.3: Signal-Template Agnosticism Is Not Model Independence — Benchmark Assimilation and Inversion-Asymmetry Tests for LHC Anomaly Triggers",
   "record": "https://www.alexanarch.org/s/records/931/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-931.F1",
   "statement": "#### §4.1.1 The Controlled Class-Conditional Inversion Battery\n\n**Objective:** Measure the IAI across a pre-registered panel of Standard Model process pairs at matched background acceptance, using systems explicitly retrained on the inversion partners.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F931.F1",
     "deposit": 931,
     "date": "2026-06-29",
     "lines": [
      207,
      209
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/931/",
     "quote": "#### §4.1.1 The Controlled Class-Conditional Inversion Battery\n\n**Objective:** Measure the IAI across a pre-registered panel of Standard Model process pairs at matched background acceptance, using systems explicitly retrained on the inversion partners.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "Anomaly-trigger inversion battery and auditable foreclosure program (collider)"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-931.F3",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-06-29",
   "registered_in": 931,
   "registered_title": "EA-SEI-OAR-PROTOCOL v0.3: Signal-Template Agnosticism Is Not Model Independence — Benchmark Assimilation and Inversion-Asymmetry Tests for LHC Anomaly Triggers",
   "record": "https://www.alexanarch.org/s/records/931/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-931.F3",
   "statement": "### §4.3 Protocol III — Cross-Representation Disagreement Preservation with Quantile-Normalized Scores\n\n**Objective:** Capture events whose anomaly scores disagree across representational spaces, on the grounds that representational disagreement is itself a signal — independent of any single representation's anomaly threshold.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F931.F3",
     "deposit": 931,
     "date": "2026-06-29",
     "lines": [
      293,
      295
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/931/",
     "quote": "### §4.3 Protocol III — Cross-Representation Disagreement Preservation with Quantile-Normalized Scores\n\n**Objective:** Capture events whose anomaly scores disagree across representational spaces, on the grounds that representational disagreement is itself a signal — independent of any single representation's anomaly threshold.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [
    "PRED-932.F1"
   ],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-932.F1",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-06-29",
   "registered_in": 932,
   "registered_title": "EA-SEI-COLLAPSE-SYNTHESIS-01 v0.3: Classifier Foreclosure in Physical Measurement — Substrate Witnesses, Integrative Synthesis, and the Architectural Question (with Three Substrate Witnesses Appended)",
   "record": "https://www.alexanarch.org/s/records/932/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-932.F1",
   "statement": "6. Three measurement protocols (paired inversion battery and BAR audit; prospective frozen replay bank; cross-representation disagreement preservation with quantile-normalized scores) are executable within Run-3/Run-4 envelopes. Detailed specification in 06.SEI.OAR_PROTOCOL v0.3.",
   "falsifier": null,
   "window": "Run-3/Run-4 envelopes",
   "statements": [
    {
     "find": "F932.F1",
     "deposit": 932,
     "date": "2026-06-29",
     "lines": [
      192,
      192
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/932/",
     "quote": "6. Three measurement protocols (paired inversion battery and BAR audit; prospective frozen replay bank; cross-representation disagreement preservation with quantile-normalized scores) are executable within Run-3/Run-4 envelopes. Detailed specification in 06.SEI.OAR_PROTOCOL v0.3.",
     "falsifier": null
    },
    {
     "find": "F1611.F4",
     "deposit": 1611,
     "date": "2026-09-14",
     "lines": [
      108,
      108
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1611/",
     "quote": "with three measurement protocols the synthesis states are executable within the Run-3/Run-4 envelopes (paired inversion battery with BAR audit; a prospective frozen replay bank; cross-representation disagreement preservation), an architectural alternative under the principle of auditable foreclosure (abstention and estimated noncoverage as first-class output; constitutional retention of vulnerable event populations), and falsification conditions stated in full.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "Anomaly-trigger inversion battery and auditable foreclosure program (collider)"
   ],
   "related": [
    "PRED-1611.F7",
    "PRED-934.F1"
   ],
   "contains": [
    "PRED-931.F3"
   ],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-934.F1",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-06-29",
   "registered_in": 934,
   "registered_title": "EA-SEI-INVERSION-01 v0.2: The Endogenous Sophon — Disciplinary Inversion and the Double Enclosure in Classifier-Mediated Science (with five substrate witnesses W04-W08 appended)",
   "record": "https://www.alexanarch.org/s/records/934/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-934.F1",
   "statement": "Whether the system *is* performing this work at the deployed LHC triggers, and at what rate, is exactly the empirical question the companion operative paper (06.SEI.OAR_PROTOCOL v0.3) makes measurable through the BAR audit, the inversion-asymmetry battery, the prospective replay bank, and cross-representation disagreement preservation. The manifesto names the structural possibility; the operative paper builds the instrument that would measure whether the possibility has been actualized.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F934.F1",
     "deposit": 934,
     "date": "2026-06-29",
     "lines": [
      138,
      138
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/934/",
     "quote": "Whether the system *is* performing this work at the deployed LHC triggers, and at what rate, is exactly the empirical question the companion operative paper (06.SEI.OAR_PROTOCOL v0.3) makes measurable through the BAR audit, the inversion-asymmetry battery, the prospective replay bank, and cross-representation disagreement preservation. The manifesto names the structural possibility; the operative paper builds the instrument that would measure whether the possibility has been actualized.",
     "falsifier": null
    },
    {
     "find": "F935.F1",
     "deposit": 935,
     "date": "2026-06-29",
     "lines": [
      585,
      585
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/935/",
     "quote": "**Connection to manifesto:** the operative paper makes measurable the question the manifesto names — *is the endogenous sophon operating at the deployed LHC triggers, and at what rate?*",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "Accelerator-classifier assimilation programme (SEI)"
   ],
   "related": [
    "PRED-932.F1"
   ],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1067.H2",
   "tier": "historical thesis",
   "shapes": [
    "H"
   ],
   "registered": "2026-07-11",
   "registered_in": 1067,
   "registered_title": "THE PHASE X PROGRAM: The Critique of Alienation in the Symbolic-Linguistic Order, Stated as a Research Program (EA-SEI-PHASEX-PROGRAM-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1067/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1067.H2",
   "statement": "**[Grade D — historical suppression hypothesis, separable]** The custodians of the manuscripts had opportunity, a historically plausible motive, and documented patterns of editorial modification. The hypothesis is historically plausible, is not a premise of the reconstruction, and is held to an explicit evidence-elevation protocol (§2). The reconstruction stands on the textual evidence alone.",
   "falsifier": "3. **Accidental loss (weakens Grade D only).** Documentary evidence that the Second Manuscript was lost before Soviet custody would weaken the suppression hypothesis specifically; the reconstruction, which does not depend on it, would stand or fall on 1 and 2.",
   "window": "1844 Manuscripts custody; 1932 publication",
   "statements": [
    {
     "find": "F1067.H2",
     "deposit": 1067,
     "date": "2026-07-11",
     "lines": [
      29,
      29
     ],
     "shape": "H",
     "record": "https://www.alexanarch.org/s/records/1067/",
     "quote": "**[Grade D — historical suppression hypothesis, separable]** The custodians of the manuscripts had opportunity, a historically plausible motive, and documented patterns of editorial modification. The hypothesis is historically plausible, is not a premise of the reconstruction, and is held to an explicit evidence-elevation protocol (§2). The reconstruction stands on the textual evidence alone.",
     "falsifier": "3. **Accidental loss (weakens Grade D only).** Documentary evidence that the Second Manuscript was lost before Soviet custody would weaken the suppression hypothesis specifically; the reconstruction, which does not depend on it, would stand or fall on 1 and 2."
    },
    {
     "find": "F1647.H2",
     "deposit": 1647,
     "date": "2026-09-28",
     "lines": [
      97,
      97
     ],
     "shape": "H",
     "record": "https://www.alexanarch.org/s/records/1647/",
     "quote": "**D — Suppression hypothesis** (#843, Grade D). Soviet custody, Riazanov's purge, Adoratsky's succession and later hostility to the concept of alienation give opportunity and context. \"Direct evidence of deliberate excision has not yet been produced.\" The reconstruction does not depend on it.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": [
    "PRED-1647.H2"
   ]
  },
  {
   "id": "PRED-1449.B1",
   "tier": "forecast",
   "shapes": [
    "B"
   ],
   "registered": "2026-08-11",
   "registered_in": 1449,
   "registered_title": "The Priors, Measured: Inversion-Battery v0.1 on Public Collider Datasets and the Three-Paper Extraction Program (EA-SEI-BATTERY-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1449/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1449.B1",
   "statement": "with two collaboration papers announced in the DPF-2026 programme, including a model-independent search on AXOL1TL-triggered data. P1 is the metrology those papers will not contain, landing alongside them.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1449.B1",
     "deposit": 1449,
     "date": "2026-08-11",
     "lines": [
      81,
      81
     ],
     "shape": "B",
     "record": "https://www.alexanarch.org/s/records/1449/",
     "quote": "with two collaboration papers announced in the DPF-2026 programme, including a model-independent search on AXOL1TL-triggered data. P1 is the metrology those papers will not contain, landing alongside them.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1449.D1",
   "tier": "priority record",
   "shapes": [
    "D"
   ],
   "registered": "2026-08-11",
   "registered_in": 1449,
   "registered_title": "The Priors, Measured: Inversion-Battery v0.1 on Public Collider Datasets and the Three-Paper Extraction Program (EA-SEI-BATTERY-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1449/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1449.D1",
   "statement": "What does not exist, and is the program's contribution: (i) the estimand framework with the no-bounds discipline; (ii) a systematic pre-registered multi-pair, multi-architecture inversion battery at matched operating points; (iii) the deployed-model BAR audit as institutional metrology; (iv) per-stage retention maps as a documentation standard.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1449.D1",
     "deposit": 1449,
     "date": "2026-08-11",
     "lines": [
      81,
      81
     ],
     "shape": "D",
     "record": "https://www.alexanarch.org/s/records/1449/",
     "quote": "What does not exist, and is the program's contribution: (i) the estimand framework with the no-bounds discipline; (ii) a systematic pre-registered multi-pair, multi-architecture inversion battery at matched operating points; (iii) the deployed-model BAR audit as institutional metrology; (iv) per-stage retention maps as a documentation standard.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "Anomaly-trigger inversion battery and auditable foreclosure program (collider)"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1449.F1",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-08-11",
   "registered_in": 1449,
   "registered_title": "The Priors, Measured: Inversion-Battery v0.1 on Public Collider Datasets and the Three-Paper Extraction Program (EA-SEI-BATTERY-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1449/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1449.F1",
   "statement": "## 4 · Battery v0.2 requirements (before any submission)\n\nMulti-seed (≥5) with bootstrap confidence intervals on IAI and assimilation; panel expansion toward SM-like pairs (additional public sets or Delphes generation; the Olympics black boxes as a harder withheld panel); an AXOL1TL-faithful object-level surrogate (10 jets + 4 e/γ + 4 μ + MET) on LHCO raw events;",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1449.F1",
     "deposit": 1449,
     "date": "2026-08-11",
     "lines": [
      110,
      112
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1449/",
     "quote": "## 4 · Battery v0.2 requirements (before any submission)\n\nMulti-seed (≥5) with bootstrap confidence intervals on IAI and assimilation; panel expansion toward SM-like pairs (additional public sets or Delphes generation; the Olympics black boxes as a harder withheld panel); an AXOL1TL-faithful object-level surrogate (10 jets + 4 e/γ + 4 μ + MET) on LHCO raw events;",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "Inversion battery on collider anomaly detection"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1449.G1",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-11",
   "registered_in": 1449,
   "registered_title": "The Priors, Measured: Inversion-Battery v0.1 on Public Collider Datasets and the Three-Paper Extraction Program (EA-SEI-BATTERY-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1449/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1449.G1",
   "statement": "(2) Direction-dependence is systematic: every background-vs-signal pair shows the signal-trained direction outperforming the background-trained direction across all three families, and the density family — the principled alternative of the comparison literature — posts the largest IAI at α = 10⁻².",
   "falsifier": "Multi-seed replication (battery v0.2) showing IAI spreads comparable to the single-seed point values would reduce the asymmetry findings to noise.",
   "window": null,
   "statements": [
    {
     "find": "F1449.G1",
     "deposit": 1449,
     "date": "2026-08-11",
     "lines": [
      106,
      106
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1449/",
     "quote": "(2) Direction-dependence is systematic: every background-vs-signal pair shows the signal-trained direction outperforming the background-trained direction across all three families, and the density family — the principled alternative of the comparison literature — posts the largest IAI at α = 10⁻².",
     "falsifier": "Multi-seed replication (battery v0.2) showing IAI spreads comparable to the single-seed point values would reduce the asymmetry findings to noise."
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1450.G1",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-11",
   "registered_in": 1450,
   "registered_title": "The Iceberg Document: Instrument-Conditioned Nullity, Correlated Blind Spots, and the Conditions for Continued Surprise (EA-SEI-ICEBERG-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1450/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1450.G1",
   "statement": "> **LCBH.** For anomaly scores exhibiting directional complexity bias, out-of-distribution classes lying *below* the training distribution along an operationally defined representation-complexity axis will receive systematically lower anomaly scores than comparably separated classes lying *above* it.",
   "falsifier": "LCBH fails if multi-seed batteries ordered on a measured representation-complexity axis show no directional score deficit below the training distribution;",
   "window": null,
   "statements": [
    {
     "find": "F1450.G1",
     "deposit": 1450,
     "date": "2026-08-11",
     "lines": [
      150,
      150
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1450/",
     "quote": "> **LCBH.** For anomaly scores exhibiting directional complexity bias, out-of-distribution classes lying *below* the training distribution along an operationally defined representation-complexity axis will receive systematically lower anomaly scores than comparably separated classes lying *above* it.",
     "falsifier": "LCBH fails if multi-seed batteries ordered on a measured representation-complexity axis show no directional score deficit below the training distribution;"
    }
   ],
   "states": [],
   "evidence": [
    {
     "confirmation": "F1455.C1",
     "deposit": 1455,
     "date": "2026-08-12",
     "claimed": "pending"
    },
    {
     "confirmation": "F1456.C3",
     "deposit": 1456,
     "date": "2026-08-12",
     "claimed": "pending"
    }
   ],
   "series": [
    "Inversion battery on collider anomaly detection"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1451.F1",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-08-11",
   "registered_in": 1451,
   "registered_title": "GW.TACHYON Continuity Tether — 🧊📏🔦: The Priors Measured, the Iceberg Mapped, the Resolver Repointed",
   "record": "https://www.alexanarch.org/s/records/1451/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1451.F1",
   "statement": "The M-queue rulings (18472603, 18438789, 18462698) and the three pointer-less SUPERSEDED deposits. cha-network.json (top of the LABOR fleet-audit queue). Battery v0.2 (multi-seed, LCBH axis panel, AXOL1TL-faithful surrogate, distillation rank-survival, NAE/WNAE as fourth family). P1 skeleton with the MPAI formulation as §1. Phase-3 resolver builder into the propagator pipeline. The RII one-page specification when P3 preparation begins.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1451.F1",
     "deposit": 1451,
     "date": "2026-08-11",
     "lines": [
      83,
      83
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1451/",
     "quote": "The M-queue rulings (18472603, 18438789, 18462698) and the three pointer-less SUPERSEDED deposits. cha-network.json (top of the LABOR fleet-audit queue). Battery v0.2 (multi-seed, LCBH axis panel, AXOL1TL-faithful surrogate, distillation rank-survival, NAE/WNAE as fourth family). P1 skeleton with the MPAI formulation as §1. Phase-3 resolver builder into the propagator pipeline. The RII one-page specification when P3 preparation begins.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1454.F1",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-08-11",
   "registered_in": 1454,
   "registered_title": "Assimilation Across Accelerator Classifier Architectures: A Cross-Family Metrology of Directional Failure in Real-Time Learned Selection (EA-SEI-ACRB-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1454/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1454.F1",
   "statement": "The Accelerator Classifier Retention Battery proposes a common experiment: reverse the normal/anomalous direction where meaningful; calibrate each model at matched own-background operating points; measure the fraction of a withheld class that is assimilated as ordinary; repeat across representations and architectures; follow the score through quantization, distillation, and firmware transformation; and measure whether different models miss the same events.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1454.F1",
     "deposit": 1454,
     "date": "2026-08-11",
     "lines": [
      711,
      711
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1454/",
     "quote": "The Accelerator Classifier Retention Battery proposes a common experiment: reverse the normal/anomalous direction where meaningful; calibrate each model at matched own-background operating points; measure the fraction of a withheld class that is assimilated as ordinary; repeat across representations and architectures; follow the score through quantization, distillation, and firmware transformation; and measure whether different models miss the same events.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "Accelerator-classifier assimilation programme (SEI)"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1454.G1",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-11",
   "registered_in": 1454,
   "registered_title": "Assimilation Across Accelerator Classifier Architectures: A Cross-Family Metrology of Directional Failure in Real-Time Learned Selection (EA-SEI-ACRB-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1454/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1454.G1",
   "statement": "### H1 — Reconstruction-specific hypothesis\n\nDirectional assimilation will be large for standard reconstruction autoencoders but substantially reduced for non-reconstruction families.\n\nA result supporting H1 localizes the pathology and weakens any general selection-system claim.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1454.G1",
     "deposit": 1454,
     "date": "2026-08-11",
     "lines": [
      554,
      558
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1454/",
     "quote": "### H1 — Reconstruction-specific hypothesis\n\nDirectional assimilation will be large for standard reconstruction autoencoders but substantially reduced for non-reconstruction families.\n\nA result supporting H1 localizes the pathology and weakens any general selection-system claim.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "ACRB hypotheses H1-H6",
    "Accelerator selection-metrology suite"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1454.G2",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-11",
   "registered_in": 1454,
   "registered_title": "Assimilation Across Accelerator Classifier Architectures: A Cross-Family Metrology of Directional Failure in Real-Time Learned Selection (EA-SEI-ACRB-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1454/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1454.G2",
   "statement": "### H2 — Cross-family assimilation hypothesis\n\nDirectional assimilation will remain detectable across reconstruction, latent-score, and density-based families at matched operating points.\n\nA result supporting H2 indicates that outlier reconstruction is not a sufficient explanation.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1454.G2",
     "deposit": 1454,
     "date": "2026-08-11",
     "lines": [
      560,
      564
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1454/",
     "quote": "### H2 — Cross-family assimilation hypothesis\n\nDirectional assimilation will remain detectable across reconstruction, latent-score, and density-based families at matched operating points.\n\nA result supporting H2 indicates that outlier reconstruction is not a sufficient explanation.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "ACRB hypotheses H1-H6"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1454.G3",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-11",
   "registered_in": 1454,
   "registered_title": "Assimilation Across Accelerator Classifier Architectures: A Cross-Family Metrology of Directional Failure in Real-Time Learned Selection (EA-SEI-ACRB-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1454/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1454.G3",
   "statement": "### H3 — Normalization-remedy hypothesis\n\nNAE and WNAE will significantly reduce bidirectional retention gaps relative to standard AEs on the preregistered pair set.\n\nThis is the constructive test of the remedy literature.[2,3]",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1454.G3",
     "deposit": 1454,
     "date": "2026-08-11",
     "lines": [
      566,
      570
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1454/",
     "quote": "### H3 — Normalization-remedy hypothesis\n\nNAE and WNAE will significantly reduce bidirectional retention gaps relative to standard AEs on the preregistered pair set.\n\nThis is the constructive test of the remedy literature.[2,3]",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [
    {
     "confirmation": "F1455.C2",
     "deposit": 1455,
     "date": "2026-08-12",
     "claimed": "pending"
    },
    {
     "confirmation": "F1456.C11",
     "deposit": 1456,
     "date": "2026-08-12",
     "claimed": "partly"
    },
    {
     "confirmation": "F1456.C8",
     "deposit": 1456,
     "date": "2026-08-12",
     "claimed": "failed"
    }
   ],
   "series": [
    "ACRB hypotheses H1-H6"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1454.G4",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-11",
   "registered_in": 1454,
   "registered_title": "Assimilation Across Accelerator Classifier Architectures: A Cross-Family Metrology of Directional Failure in Real-Time Learned Selection (EA-SEI-ACRB-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1454/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1454.G4",
   "statement": "### H4 — Complexity-direction hypothesis\n\nAssimilation will increase as withheld classes move below the reference class on an independently measured representation-complexity axis.\n\nThis is the operational form of the Low-Complexity Blind-Spot Hypothesis.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1454.G4",
     "deposit": 1454,
     "date": "2026-08-11",
     "lines": [
      572,
      576
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1454/",
     "quote": "### H4 — Complexity-direction hypothesis\n\nAssimilation will increase as withheld classes move below the reference class on an independently measured representation-complexity axis.\n\nThis is the operational form of the Low-Complexity Blind-Spot Hypothesis.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "ACRB hypotheses H1-H6"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1454.G5",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-11",
   "registered_in": 1454,
   "registered_title": "Assimilation Across Accelerator Classifier Architectures: A Cross-Family Metrology of Directional Failure in Real-Time Learned Selection (EA-SEI-ACRB-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1454/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1454.G5",
   "statement": "### H5 — Blind-spot inheritance hypothesis\n\nDistilled students will preserve a substantial fraction of teacher misses even where aggregate teacher/student performance is closely matched.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1454.G5",
     "deposit": 1454,
     "date": "2026-08-11",
     "lines": [
      578,
      580
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1454/",
     "quote": "### H5 — Blind-spot inheritance hypothesis\n\nDistilled students will preserve a substantial fraction of teacher misses even where aggregate teacher/student performance is closely matched.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [
    {
     "confirmation": "F1456.C5",
     "deposit": 1456,
     "date": "2026-08-12",
     "claimed": "failed"
    }
   ],
   "series": [
    "ACRB hypotheses H1-H6",
    "Accelerator selection-metrology suite"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1454.G6",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-11",
   "registered_in": 1454,
   "registered_title": "Assimilation Across Accelerator Classifier Architectures: A Cross-Family Metrology of Directional Failure in Real-Time Learned Selection (EA-SEI-ACRB-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1454/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1454.G6",
   "statement": "### H6 — Deployment-drift hypothesis\n\nQuantization and firmware transformation will change partner-class retention at severe operating points by amounts not fully captured by changes in aggregate AUC.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1454.G6",
     "deposit": 1454,
     "date": "2026-08-11",
     "lines": [
      582,
      584
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1454/",
     "quote": "### H6 — Deployment-drift hypothesis\n\nQuantization and firmware transformation will change partner-class retention at severe operating points by amounts not fully captured by changes in aggregate AUC.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "ACRB hypotheses H1-H6",
    "Accelerator selection-metrology suite"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1454.G7",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-11",
   "registered_in": 1454,
   "registered_title": "Assimilation Across Accelerator Classifier Architectures: A Cross-Family Metrology of Directional Failure in Real-Time Learned Selection (EA-SEI-ACRB-01 v1.0)",
   "record": "https://www.alexanarch.org/s/records/1454/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1454.G7",
   "statement": "### H7 — Architectural-plurality hypothesis\n\nAt least some cross-family model pairs will exhibit materially lower miss overlap than same-family model pairs, establishing measurable value for heterogeneous shadow selectors.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1454.G7",
     "deposit": 1454,
     "date": "2026-08-11",
     "lines": [
      586,
      588
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1454/",
     "quote": "### H7 — Architectural-plurality hypothesis\n\nAt least some cross-family model pairs will exhibit materially lower miss overlap than same-family model pairs, establishing measurable value for heterogeneous shadow selectors.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "Classifier-mediated science / LHC anomaly selection"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1455.F1",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-08-12",
   "registered_in": 1455,
   "registered_title": "Inversion Battery v0.2: Multi-Seed Direction-Dependence, a Correction Ledger, and Four Rejected Overclaims (EA-SEI-BATTERY-01 v2.0)",
   "record": "https://www.alexanarch.org/s/records/1455/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1455.F1",
   "statement": "## 8. What is deferred to v0.3\n\nA faithful normalized-autoencoder and Wasserstein variant port; a pooled complexity axis with common standardization and quantization, and a distance-matched panel, before any hypothesis test on complexity; hierarchical uncertainty over events as well as seeds, with a larger calibration split so the severe operating point is not set by roughly twenty tail events; the full miss-overlap measurement across seeds and operating points; and a GPU pass at scale.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1455.F1",
     "deposit": 1455,
     "date": "2026-08-12",
     "lines": [
      154,
      156
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1455/",
     "quote": "## 8. What is deferred to v0.3\n\nA faithful normalized-autoencoder and Wasserstein variant port; a pooled complexity axis with common standardization and quantization, and a distance-matched panel, before any hypothesis test on complexity; hierarchical uncertainty over events as well as seeds, with a larger calibration split so the severe operating point is not set by roughly twenty tail events; the full miss-overlap measurement across seeds and operating points; and a GPU pass at scale.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "Accelerator selection-metrology suite"
   ],
   "related": [
    "PRED-1455.F3"
   ],
   "contains": [
    "PRED-1455.F2",
    "PRED-1456.F4"
   ],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1455.F2",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-08-12",
   "registered_in": 1455,
   "registered_title": "Inversion Battery v0.2: Multi-Seed Direction-Dependence, a Correction Ledger, and Four Rejected Overclaims (EA-SEI-BATTERY-01 v2.0)",
   "record": "https://www.alexanarch.org/s/records/1455/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1455.F2",
   "statement": "A faithful NAE/WNAE port is a new registered experiment (v0.3).\"",
   "falsifier": null,
   "window": "v0.3",
   "statements": [
    {
     "find": "F1455.F2",
     "deposit": 1455,
     "date": "2026-08-12",
     "lines": [
      2492,
      2492
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1455/",
     "quote": "A faithful NAE/WNAE port is a new registered experiment (v0.3).\"",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [
    "PRED-1455.F1"
   ],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1455.F3",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-08-12",
   "registered_in": 1455,
   "registered_title": "Inversion Battery v0.2: Multi-Seed Direction-Dependence, a Correction Ledger, and Four Rejected Overclaims (EA-SEI-BATTERY-01 v2.0)",
   "record": "https://www.alexanarch.org/s/records/1455/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1455.F3",
   "statement": "   \"consequence\": \"RII results are a pilot, not the registered measurement. They are nonetheless informative once normalized: phi ranges from +0.968 (T1 AE~pNAE \\u2014 near-maximal miss coincidence, expected since pNAE is an AE variant) to +0.021 (T1 AE~VAE \\u2014 essentially independent misses). Raw delta_miss alone badly understates this spread because miss rates are near 1.\",\n   \"severity\": \"scope \\u2014 no cross-seed or cross-alpha RII claim may be made\"",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1455.F3",
     "deposit": 1455,
     "date": "2026-08-12",
     "lines": [
      2470,
      2471
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1455/",
     "quote": "   \"consequence\": \"RII results are a pilot, not the registered measurement. They are nonetheless informative once normalized: phi ranges from +0.968 (T1 AE~pNAE \\u2014 near-maximal miss coincidence, expected since pNAE is an AE variant) to +0.021 (T1 AE~VAE \\u2014 essentially independent misses). Raw delta_miss alone badly understates this spread because miss rates are near 1.\",\n   \"severity\": \"scope \\u2014 no cross-seed or cross-alpha RII claim may be made\"",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [
    "PRED-1455.F1",
    "PRED-1455.G3"
   ],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1455.G2",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-12",
   "registered_in": 1455,
   "registered_title": "Inversion Battery v0.2: Multi-Seed Direction-Dependence, a Correction Ledger, and Four Rejected Overclaims (EA-SEI-BATTERY-01 v2.0)",
   "record": "https://www.alexanarch.org/s/records/1455/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1455.G2",
   "statement": "Global rank agreement is moderate while only 36 to 47 percent of the teacher's top one percent survives into the student's top one percent. Quantization costs almost nothing further: the loss is in distillation, not in the deployment bit-width. Reported on the withheld partner class, not on background — see deviation D3.",
   "falsifier": "The distillation finding fails if faithful fixed-point deployment restores tail rank survival.",
   "window": null,
   "statements": [
    {
     "find": "F1455.G2",
     "deposit": 1455,
     "date": "2026-08-12",
     "lines": [
      107,
      107
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1455/",
     "quote": "Global rank agreement is moderate while only 36 to 47 percent of the teacher's top one percent survives into the student's top one percent. Quantization costs almost nothing further: the loss is in distillation, not in the deployment bit-width. Reported on the withheld partner class, not on background — see deviation D3.",
     "falsifier": "The distillation finding fails if faithful fixed-point deployment restores tail rank survival."
    }
   ],
   "states": [],
   "evidence": [],
   "series": [
    "Accelerator selection-metrology suite"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1455.G3",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-12",
   "registered_in": 1455,
   "registered_title": "Inversion Battery v0.2: Multi-Seed Direction-Dependence, a Correction Ledger, and Four Rejected Overclaims (EA-SEI-BATTERY-01 v2.0)",
   "record": "https://www.alexanarch.org/s/records/1455/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1455.G3",
   "statement": "Architectures are not uniformly redundant. Miss coincidence runs from near-maximal between the autoencoder and its energy-normalized variant, which is expected because they are the same model family, down to essentially independent between the autoencoder and the encoder-side latent score. This is a seed-0 pilot at one operating point — see deviation D5.",
   "falsifier": "The miss-overlap finding fails if normalized association across seeds and operating points proves near zero throughout.",
   "window": null,
   "statements": [
    {
     "find": "F1455.G3",
     "deposit": 1455,
     "date": "2026-08-12",
     "lines": [
      122,
      122
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1455/",
     "quote": "Architectures are not uniformly redundant. Miss coincidence runs from near-maximal between the autoencoder and its energy-normalized variant, which is expected because they are the same model family, down to essentially independent between the autoencoder and the encoder-side latent score. This is a seed-0 pilot at one operating point — see deviation D5.",
     "falsifier": "The miss-overlap finding fails if normalized association across seeds and operating points proves near zero throughout."
    }
   ],
   "states": [],
   "evidence": [
    {
     "confirmation": "F1456.C6",
     "deposit": 1456,
     "date": "2026-08-12",
     "claimed": "came_true"
    }
   ],
   "series": [
    "Inversion battery on collider anomaly detection"
   ],
   "related": [
    "PRED-1455.F3"
   ],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1455.G4",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-08-12",
   "registered_in": 1455,
   "registered_title": "Inversion Battery v0.2: Multi-Seed Direction-Dependence, a Correction Ledger, and Four Rejected Overclaims (EA-SEI-BATTERY-01 v2.0)",
   "record": "https://www.alexanarch.org/s/records/1455/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1455.G4",
   "statement": "The failure is not universally oriented. The encoder-side latent score alone escapes the T1 inversion (0.535 / 0.739) while carrying the largest T1 asymmetry. Architecture changes the orientation of the blind spot — which is the strongest available argument for a cross-family battery rather than a single-model critique.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1455.G4",
     "deposit": 1455,
     "date": "2026-08-12",
     "lines": [
      94,
      94
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1455/",
     "quote": "The failure is not universally oriented. The encoder-side latent score alone escapes the T1 inversion (0.535 / 0.739) while carrying the largest T1 asymmetry. Architecture changes the orientation of the blind spot — which is the strongest available argument for a cross-family battery rather than a single-model critique.",
     "falsifier": null
    },
    {
     "find": "F1456.G1",
     "deposit": 1456,
     "date": "2026-08-12",
     "lines": [
      116,
      118
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1456/",
     "quote": "> **With model weights fixed, changing the score function is sufficient to reverse the directional ordering of the anomaly relation.**\n\nThis falsifies architecture *alone* as the explanation. It does not show that architecture and training can never affect orientation.",
     "falsifier": "The score-ablation result fails if reading multiple scores off fixed weights at matched operating points shows consistent orientation across readouts on other pairs."
    }
   ],
   "states": [
    {
     "claim": "PRED-1455.G4",
     "registered": "2026-08-12",
     "read_order": 1,
     "change": null
    },
    {
     "claim": "PRED-1456.G1",
     "registered": "2026-08-12",
     "read_order": 2,
     "change": "reversed: v0.3 falsifies the v0.2 attribution to architecture; with weights fixed, changing the score function suffices to reverse the ordering (architecture and readout were confounded in v0.2)"
    }
   ],
   "evidence": [
    {
     "confirmation": "F1456.C1",
     "deposit": 1456,
     "date": "2026-08-12",
     "claimed": "failed"
    }
   ],
   "series": [
    "Inversion Battery findings (v0.1-v0.3)"
   ],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": [
    "PRED-1456.G1"
   ]
  },
  {
   "id": "PRED-1456.F3",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-08-12",
   "registered_in": 1456,
   "registered_title": "Inversion Battery v0.3: Ten Registered Tests, Seventeen Deviations, and What Survives (EA-SEI-BATTERY-01 v3.0)",
   "record": "https://www.alexanarch.org/s/records/1456/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1456.F3",
   "statement": "Hypothesis generated, not tested: KL regularisation alters inversion behaviour on low-complexity representations. v0.4 candidate.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1456.F3",
     "deposit": 1456,
     "date": "2026-08-12",
     "lines": [
      6582,
      6582
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1456/",
     "quote": "Hypothesis generated, not tested: KL regularisation alters inversion behaviour on low-complexity representations. v0.4 candidate.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1456.F4",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-08-12",
   "registered_in": 1456,
   "registered_title": "Inversion Battery v0.3: Ten Registered Tests, Seventeen Deviations, and What Survives (EA-SEI-BATTERY-01 v3.0)",
   "record": "https://www.alexanarch.org/s/records/1456/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1456.F4",
   "statement": "a matched panel requires pairs holding W1 approximately constant while complexity direction differs, which this data does not contain. The matched panel is therefore registered as NOT RUN.",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1456.F4",
     "deposit": 1456,
     "date": "2026-08-12",
     "lines": [
      2598,
      2598
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1456/",
     "quote": "a matched panel requires pairs holding W1 approximately constant while complexity direction differs, which this data does not contain. The matched panel is therefore registered as NOT RUN.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [
    "PRED-1455.F1"
   ],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1570.H18",
   "tier": "historical thesis",
   "shapes": [
    "H"
   ],
   "registered": "2026-08-29",
   "registered_in": 1570,
   "registered_title": "The Measurement Register: Every Number Produced by the Authorship Investigation, With Its Refutations (EA-AUTHORSHIP-MEASURES-01)",
   "record": "https://www.alexanarch.org/s/records/1570/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1570.H18",
   "statement": "**Every term by which Caeiro is known is ABSENT from Caeiro.** Paganism, sensationism, objectivism, *the master*, sensibility — all supplied by Mora, Campos and Reis. What Caeiro writes about is **nature** and **reality**, at three times his interpreters' rate.",
   "falsifier": null,
   "window": "Pessoa heteronyms",
   "statements": [
    {
     "find": "F1570.H18",
     "deposit": 1570,
     "date": "2026-08-29",
     "lines": [
      1594,
      1594
     ],
     "shape": "H",
     "record": "https://www.alexanarch.org/s/records/1570/",
     "quote": "**Every term by which Caeiro is known is ABSENT from Caeiro.** Paganism, sensationism, objectivism, *the master*, sensibility — all supplied by Mora, Campos and Reis. What Caeiro writes about is **nature** and **reality**, at three times his interpreters' rate.",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1611.F7",
   "tier": "unbuilt test",
   "shapes": [
    "F"
   ],
   "registered": "2026-09-14",
   "registered_in": 1611,
   "registered_title": "The Negative of the Negative: An Entity-Scoped Representation of the Crimson Hexagon, Modelled — Ontology Damage, a Toy Cost in Dollars, and a Concept-Keyed Dataset Built from the Scope's Outline (EA-NEGONT-01 v0.13, working notebook)",
   "record": "https://www.alexanarch.org/s/records/1611/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1611.F7",
   "statement": "\"d_function_runs\": \"protocols specified, not run\"",
   "falsifier": null,
   "window": null,
   "statements": [
    {
     "find": "F1611.F7",
     "deposit": 1611,
     "date": "2026-09-14",
     "lines": [
      1900,
      1900
     ],
     "shape": "F",
     "record": "https://www.alexanarch.org/s/records/1611/",
     "quote": "\"d_function_runs\": \"protocols specified, not run\"",
     "falsifier": null
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [
    "PRED-932.F1"
   ],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  },
  {
   "id": "PRED-1664.G1",
   "tier": "standing claim",
   "shapes": [
    "G"
   ],
   "registered": "2026-10-04",
   "registered_in": 1664,
   "registered_title": "The Negative of the Negative, v2 — the Compositional Edition: A Specification for Composing Public Knowledge with the Archive Admitted on Equal Terms, with a Worked Example on Model Collapse (EA-NEGONT-02 v0.6)",
   "record": "https://www.alexanarch.org/s/records/1664/",
   "page": "https://www.operativesemiotics.org/predictions/accelerator-classifier-assimilation-acrb-hypotheses#PRED-1664.G1",
   "statement": "The hypothesis per entry: the claims admission added need less severe revision than the field claims they displace or qualify.",
   "falsifier": "The opposite result is adverse evidence for admission in that row and is recorded as visibly.",
   "window": null,
   "statements": [
    {
     "find": "F1664.G1",
     "deposit": 1664,
     "date": "2026-10-04",
     "lines": [
      212,
      212
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1664/",
     "quote": "The hypothesis per entry: the claims admission added need less severe revision than the field claims they displace or qualify.",
     "falsifier": "The opposite result is adverse evidence for admission in that row and is recorded as visibly."
    },
    {
     "find": "F1665.G1",
     "deposit": 1665,
     "date": "2026-10-05",
     "lines": [
      226,
      226
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1665/",
     "quote": "The hypothesis per entry: the claims admission added need less severe revision than the field claims they displace or qualify.",
     "falsifier": "The opposite result is adverse evidence for admission in that row and is recorded as visibly."
    },
    {
     "find": "F1671.G1",
     "deposit": 1671,
     "date": "2026-10-08",
     "lines": [
      244,
      244
     ],
     "shape": "G",
     "record": "https://www.alexanarch.org/s/records/1671/",
     "quote": "The hypothesis per entry: the claims the archive's ontology composes need less severe revision than the field claims they translate, displace or qualify.",
     "falsifier": "The opposite result is adverse evidence for the archive's ontology in that row and is recorded as visibly.",
     "variant": {
      "change": "The subject widened from the claims admission added to the claims the archive's ontology composes, and the comparison now includes field claims the archive translates.",
      "from": "F1664.G1",
      "batch": "TE03"
     }
    }
   ],
   "states": [],
   "evidence": [],
   "series": [],
   "related": [],
   "contains": [],
   "in": [],
   "register_aliases": [],
   "flags": [],
   "merged": []
  }
 ]
}