{
  "n_prompts": 1000,
  "invalid": 0,
  "read_positions": 97419,
  "positions_per_prompt": 97.419,
  "bands": 5,
  "classes": [
    "keyword",
    "identifier",
    "symbol",
    "newline-indent"
  ],
  "histogram": {
    "b0:keyword": 2054,
    "b0:identifier": 10144,
    "b0:symbol": 6035,
    "b0:newline-indent": 1647,
    "b1:keyword": 1526,
    "b1:identifier": 9275,
    "b1:symbol": 6501,
    "b1:newline-indent": 2181,
    "b2:keyword": 1416,
    "b2:identifier": 9630,
    "b2:symbol": 5485,
    "b2:newline-indent": 2942,
    "b3:keyword": 1271,
    "b3:identifier": 9826,
    "b3:symbol": 4906,
    "b3:newline-indent": 3480,
    "b4:keyword": 982,
    "b4:identifier": 9102,
    "b4:symbol": 5220,
    "b4:newline-indent": 3796
  },
  "thinnest_cell": "b4:keyword",
  "size_for_threshold": {
    "100": 250,
    "200": 250,
    "400": 500,
    "800": 1000
  },
  "note": "C is recorded at every position, so one prompt contributes as many yardstick observations as it has tokens"
}