{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "9a59273e",
   "metadata": {},
   "source": [
    "# Declaration and persistent state change — completed secondary analysis\n",
    "\n",
    "**Version 1.0 | 17 September 2026 | Prepared for Joseph Anady**\n",
    "\n",
    "This notebook reproduces numerical findings from verified **real-data derived arrays**, selected actual-checkpoint metrics, and complete sensitivity outcomes. Original tensor processing occurred in a source-connected session. The full raw tensors are not bundled here and are **not** loaded by this notebook. Use `code/checkpoint_analysis.py --download` for the full pinned-source computation.\n",
    "\n",
    "Scope: **one training run**, 167 saved adapter checkpoints, 792 gradient observations, and 79 evaluation losses. No new model inference, training, declaration measurement, or consciousness classification is performed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "abffee15",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "import sys, json, subprocess, runpy\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "ROOT = Path.cwd()\n",
    "assert (ROOT / \"provenance/source_verification.json\").is_file(), \"Run from the package root.\"\n",
    "sys.path.insert(0, str(ROOT / \"code\"))\n",
    "print(\"Python\", sys.version.split()[0])\n",
    "print(\"NumPy\", np.__version__)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ae423ee9",
   "metadata": {},
   "source": [
    "## 1. Verify transferred artifacts against source-session receipts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7d73f0f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "result = subprocess.run([sys.executable, \"code/verify_package.py\"], cwd=ROOT,\n",
    "                        text=True, capture_output=True, check=True)\n",
    "print(result.stdout)\n",
    "provenance = json.loads((ROOT / \"provenance/source_verification.json\").read_text())\n",
    "print(json.dumps(provenance, indent=2))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cb2277db",
   "metadata": {},
   "source": [
    "## 2. Recompute full-log summaries\n",
    "\n",
    "The two-line descriptive split deliberately retains the original 300-step pilot domain. No inference treats sequential observations as independent trials."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a2c215cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "from analyze_tables import main\n",
    "summary = main()\n",
    "print(json.dumps(summary, indent=2))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dfd8b477",
   "metadata": {},
   "source": [
    "## 3. Inspect actual-checkpoint summaries and all sensitivity conditions\n",
    "\n",
    "The selected table contains 16 observed checkpoints. It must not be mistaken for all 167 snapshots. All 18 lag/eligibility settings are retained below. The source-style amplitude filter is factor-based and is not invariant under arbitrary reciprocal factor rescaling, even though the effective-update corner is invariant."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ce8484ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "selected = pd.read_csv(ROOT / \"data/selected_checkpoint_metrics.csv\", float_precision=\"round_trip\")\n",
    "sensitivity = pd.read_csv(ROOT / \"data/corner_sensitivity.csv\", float_precision=\"round_trip\")\n",
    "print(selected.to_string(index=False))\n",
    "print(\"\\nComplete sensitivity table:\")\n",
    "print(sensitivity.to_string(index=False))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "be5729ba",
   "metadata": {},
   "outputs": [],
   "source": [
    "comparisons = json.loads((ROOT / \"data/checkpoint_comparisons.json\").read_text())\n",
    "validation = json.loads((ROOT / \"data/tensor_validation.json\").read_text())\n",
    "print(json.dumps(comparisons, indent=2))\n",
    "print(json.dumps(validation, indent=2))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "350f3f1c",
   "metadata": {},
   "source": [
    "## 4. Computational tests\n",
    "\n",
    "These check data identity, numerical properties, and mathematical implementations. Synthetic matrix tests validate formulas; they are not simulated consciousness experiments."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c95bb845",
   "metadata": {},
   "outputs": [],
   "source": [
    "test = subprocess.run([sys.executable, \"-m\", \"unittest\", \"discover\", \"-s\", \"code\", \"-p\", \"test_*.py\", \"-v\"],\n",
    "                      cwd=ROOT, text=True, capture_output=True)\n",
    "print(test.stdout + test.stderr)\n",
    "assert test.returncode == 0\n",
    "(ROOT / \"results/tests.txt\").write_text(test.stdout + test.stderr)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0fc184f",
   "metadata": {},
   "source": [
    "## 5. Figures\n",
    "\n",
    "Each plot is separate. Full-series versus selected-checkpoint scope is identified in titles and manuscript captions. Reference lines identify calculations, not declarations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ee3ab451",
   "metadata": {},
   "outputs": [],
   "source": [
    "from make_figures import figures\n",
    "figures(show=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1646f33f",
   "metadata": {},
   "source": [
    "## Interpretation boundary\n",
    "\n",
    "The first nonzero update (step 2), gradient maximum (169), descriptive prefix split (after 181), and filtered local-bend maximum (190) are **different quantities**. Without the filter, the shorter local-bend windows select 725. This makes measurement specification central to any proposed onset claim.\n",
    "\n",
    "The acquired artifacts contain no declaration variable. This notebook therefore cannot estimate declaration-before-change ordering, within-chat persistence, an onset distribution across independent runs, or subjective consciousness. The manuscript treats those as distinct hypotheses requiring additional observations.\n",
    "\n",
    "## Full source-connected reproduction\n",
    "\n",
    "```bash\n",
    "python code/checkpoint_analysis.py --download --raw-dir raw_sources --output-dir recomputed\n",
    "python code/verify_recomputed.py recomputed raw_sources\n",
    "```\n",
    "\n",
    "Those commands fetch public pinned JSON and safetensors, compute the full geometry, and verify outputs. They do not load the full base model, train a model, or make paid inference requests."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "name": "python",
   "version": "3.13.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}