{
  "capturedAt": "2026-10-04T08:41:44.593Z",
  "url": "https://m-sanchez-calibration-explorer.static.hf.space/index.html#workspace",
  "spaceCommit": "6b7e39ed9b73e6d292e9df1510f96a04336ff0f7",
  "sourceCommit": "b4d3107c008c9b39828254035d936764e58902f3",
  "appVersion": "0.2.0",
  "sourceSha256": "9ce40339840b8e81d3c709af90b81d4e9c5108b53872625f0f932cfaebc0d05f",
  "previewMatchesActualDownload": true,
  "operator": "Codex agent",
  "independentHumanParticipants": 0,
  "captureType": "Sequential screenshots of actual UI actions, not a real-time video recording",
  "title": "Calibration Explorer 0.2.0 reference workflow",
  "context": "Captured on 4 October 2026 from the public version 0.2.0 using the UCI handwritten-digit reference. These are Codex-operated engineering observations, with zero independent human participants. The 80% threshold is illustrative, not optimised or recommended.",
  "steps": [
    {
      "file": "01-reference.png",
      "title": "Load the reference",
      "caption": "Load the handwritten-digit reference. Policy validation is selected; the 1,797 test rows remain unopened.",
      "alt": "Calibration Explorer with the UCI handwritten-digit reference loaded, Policy validation selected and the test split unopened.",
      "durationSeconds": 6.0,
      "startSeconds": 0,
      "sha256": "aa5c67f15f06cdb951f8e049a68aaf6ce8c99bd5add628abf5f3293e2382466e",
      "width": 1425,
      "height": 950
    },
    {
      "file": "02-linked-bin.png",
      "title": "Inspect a linked reliability bin",
      "caption": "Select original bin 15: 412 predictions, 98.3% mean confidence and 100% observed accuracy.",
      "alt": "Original reliability chart with bin 15 selected and linked details showing 412 predictions, 98.3% confidence and 100% accuracy.",
      "durationSeconds": 6.0,
      "startSeconds": 6.0,
      "sha256": "6582e14df612b72a9f1ef1241203d7cec3c6456bb1d0eea1a60ce5c96186d8f8",
      "width": 1425,
      "height": 950
    },
    {
      "file": "03-calibration.png",
      "title": "Fit on calibration rows",
      "caption": "Fit temperature using only 574 calibration rows. The fitted value is 0.72598.",
      "alt": "Calibrate stage showing a fitted temperature of 0.72598 from 574 calibration observations, with Policy validation still selected.",
      "durationSeconds": 6.0,
      "startSeconds": 12.0,
      "sha256": "2d8f0d6780e504f99a92ccb1d7143c677849d81eef0257b53b27ff711c884de6",
      "width": 1425,
      "height": 950
    },
    {
      "file": "04-policy-lock.png",
      "title": "Lock an illustrative threshold",
      "caption": "Lock an illustrative 80% threshold on policy validation before opening the test split.",
      "alt": "Decide stage with an 80% threshold locked at temperature 0.7260 on policy validation; the next action is Review test results.",
      "durationSeconds": 6.0,
      "startSeconds": 18.0,
      "sha256": "67a40748fb027cfb4321b75cd802796b0ab039a4e277c38ed2cec29c1a80a1c8",
      "width": 1425,
      "height": 950
    },
    {
      "file": "05-test-review.png",
      "title": "Review the test comparison",
      "caption": "Review 1,797 test predictions. Accuracy stays at 94.9%; NLL changes from 0.1760 to 0.1575.",
      "alt": "Test comparison showing 1,797 observations, unchanged 94.9% accuracy, ECE changing from 0.0365 to 0.0101 and NLL from 0.1760 to 0.1575.",
      "durationSeconds": 6.0,
      "startSeconds": 24.0,
      "sha256": "ad79416a18b468408283df1dcbd913cc15c1aaefbedc1a5138caa2b3fe99ba62",
      "width": 1425,
      "height": 950
    },
    {
      "file": "06-export-preview.png",
      "title": "Preview and save the report",
      "caption": "Preview the test report. Source rows are excluded; the actual HTML download matches this preview.",
      "alt": "Export stage with HTML, JSON and CSV controls, source rows unchecked, and a preview labelled as the same snapshot used by the HTML download.",
      "durationSeconds": 6.0,
      "startSeconds": 30.0,
      "sha256": "cd69889c107e8226db6a15b9828a8f40301667c05ca1e0822178fbdb9e141029",
      "width": 1425,
      "height": 950
    }
  ],
  "downloadedEvidence": {
    "file": "released-reference-assessment.html",
    "sha256": "6e6046fa746f4159dc10c98d3b898b0d0ad2d3b6e29edee9ea127125817c41ac",
    "previewMatchesActualDownload": true
  },
  "sequence": {
    "file": "reference-workflow-step-by-step.webm",
    "format": "WebM / VP8",
    "audio": false,
    "framesPerSecond": 10,
    "frames": 360,
    "durationSeconds": 36.0,
    "width": 1426,
    "height": 950,
    "processing": "Screenshots held sequentially without cropping, scaling, added UI, pointer movement or transitions. At most one white pixel is added on the right or bottom for even video dimensions.",
    "sha256": "9a3287fe896f505b0ba4b95d259afea8c80c0a22243a088fa7a05f3abd775f51",
    "bytes": 3079632
  }
}