{
 "cols": [
  "v1.3",
  "v1.2",
  "v1.0",
  "Jev 1.13",
  "Clef",
  "Clef-Flash",
  "Laya"
 ],
 "rows": [
  {
   "name": "AG News",
   "n": 300,
   "correct": {
    "v1.3": 271,
    "v1.2": 249,
    "v1.0": 250,
    "Jev 1.13": 248,
    "Clef": 261,
    "Clef-Flash": 266,
    "Laya": 275
   },
   "pct": {
    "v1.3": 90.3,
    "v1.2": 83.0,
    "v1.0": 83.3,
    "Jev 1.13": 82.7,
    "Clef": 87.0,
    "Clef-Flash": 88.7,
    "Laya": 91.7
   }
  },
  {
   "name": "ARC-Easy",
   "n": 300,
   "correct": {
    "v1.3": 295,
    "v1.2": 287,
    "v1.0": 285,
    "Jev 1.13": 297,
    "Clef": 299,
    "Clef-Flash": 299,
    "Laya": 144
   },
   "pct": {
    "v1.3": 98.3,
    "v1.2": 95.7,
    "v1.0": 95.0,
    "Jev 1.13": 99.0,
    "Clef": 99.7,
    "Clef-Flash": 99.7,
    "Laya": 48.0
   }
  },
  {
   "name": "BoolQ",
   "n": 300,
   "correct": {
    "v1.3": 273,
    "v1.2": 261,
    "v1.0": 262,
    "Jev 1.13": 276,
    "Clef": 278,
    "Clef-Flash": 275,
    "Laya": 242
   },
   "pct": {
    "v1.3": 91.0,
    "v1.2": 87.0,
    "v1.0": 87.3,
    "Jev 1.13": 92.0,
    "Clef": 92.7,
    "Clef-Flash": 91.7,
    "Laya": 80.7
   }
  },
  {
   "name": "HellaSwag",
   "n": 300,
   "correct": {
    "v1.3": 281,
    "v1.2": 242,
    "v1.0": 241,
    "Jev 1.13": 285,
    "Clef": 294,
    "Clef-Flash": 296,
    "Laya": 85
   },
   "pct": {
    "v1.3": 93.7,
    "v1.2": 80.7,
    "v1.0": 80.3,
    "Jev 1.13": 95.0,
    "Clef": 98.0,
    "Clef-Flash": 98.7,
    "Laya": 28.3
   }
  },
  {
   "name": "PIQA",
   "n": 300,
   "correct": {
    "v1.3": 264,
    "v1.2": 234,
    "v1.0": 246,
    "Jev 1.13": 287,
    "Clef": 273,
    "Clef-Flash": 271,
    "Laya": 161
   },
   "pct": {
    "v1.3": 88.0,
    "v1.2": 78.0,
    "v1.0": 82.0,
    "Jev 1.13": 95.7,
    "Clef": 91.0,
    "Clef-Flash": 90.3,
    "Laya": 53.7
   }
  },
  {
   "name": "PubMedQA",
   "n": 300,
   "correct": {
    "v1.3": 243,
    "v1.2": 211,
    "v1.0": 202,
    "Jev 1.13": 231,
    "Clef": 233,
    "Clef-Flash": 239,
    "Laya": 171
   },
   "pct": {
    "v1.3": 81.0,
    "v1.2": 70.3,
    "v1.0": 67.3,
    "Jev 1.13": 77.0,
    "Clef": 77.7,
    "Clef-Flash": 79.7,
    "Laya": 57.0
   }
  },
  {
   "name": "RTE",
   "n": 277,
   "correct": {
    "v1.3": 261,
    "v1.2": 246,
    "v1.0": 247,
    "Jev 1.13": 254,
    "Clef": 254,
    "Clef-Flash": 252,
    "Laya": 230
   },
   "pct": {
    "v1.3": 94.2,
    "v1.2": 88.8,
    "v1.0": 89.2,
    "Jev 1.13": 91.7,
    "Clef": 91.7,
    "Clef-Flash": 91.0,
    "Laya": 83.0
   }
  },
  {
   "name": "SST-2",
   "n": 300,
   "correct": {
    "v1.3": 286,
    "v1.2": 281,
    "v1.0": 282,
    "Jev 1.13": 286,
    "Clef": 283,
    "Clef-Flash": 289,
    "Laya": 273
   },
   "pct": {
    "v1.3": 95.3,
    "v1.2": 93.7,
    "v1.0": 94.0,
    "Jev 1.13": 95.3,
    "Clef": 94.3,
    "Clef-Flash": 96.3,
    "Laya": 91.0
   }
  },
  {
   "name": "TweetEval offensive",
   "n": 300,
   "correct": {
    "v1.3": 259,
    "v1.2": 245,
    "v1.0": 248,
    "Jev 1.13": 247,
    "Clef": 248,
    "Clef-Flash": 253,
    "Laya": 235
   },
   "pct": {
    "v1.3": 86.3,
    "v1.2": 81.7,
    "v1.0": 82.7,
    "Jev 1.13": 82.3,
    "Clef": 82.7,
    "Clef-Flash": 84.3,
    "Laya": 78.3
   }
  },
  {
   "name": "WinoGrande",
   "n": 300,
   "correct": {
    "v1.3": 273,
    "v1.2": 216,
    "v1.0": 212,
    "Jev 1.13": 268,
    "Clef": 281,
    "Clef-Flash": 294,
    "Laya": 151
   },
   "pct": {
    "v1.3": 91.0,
    "v1.2": 72.0,
    "v1.0": 70.7,
    "Jev 1.13": 89.3,
    "Clef": 93.7,
    "Clef-Flash": 98.0,
    "Laya": 50.3
   }
  },
  {
   "name": "SNLI",
   "n": 150,
   "correct": {
    "v1.3": 128,
    "v1.2": 126,
    "v1.0": 95,
    "Jev 1.13": 113,
    "Clef": 124,
    "Clef-Flash": 124,
    "Laya": 69
   },
   "pct": {
    "v1.3": 85.3,
    "v1.2": 84.0,
    "v1.0": 63.3,
    "Jev 1.13": 75.3,
    "Clef": 82.7,
    "Clef-Flash": 82.7,
    "Laya": 46.0
   }
  },
  {
   "name": "HelpSteer2 (0–4 score)",
   "n": 300,
   "pct": {
    "v1.3": 42.3,
    "v1.2": 41.3,
    "v1.0": 40.0,
    "Jev 1.13": 41.5,
    "Clef": 41.3,
    "Clef-Flash": 42.7,
    "Laya": 35.3
   },
   "note": "peers: mean over 5 option orders (1,500 decisions); ours one ask per item, over-limit requests count wrong"
  },
  {
   "name": "Banking77 (77-way)†",
   "n": 300,
   "pct": {
    "v1.3": 88.1,
    "v1.2": 80.5,
    "v1.0": null,
    "Jev 1.13": 80.0,
    "Clef": 94.6,
    "Clef-Flash": 94.7,
    "Laya": 38.7
   },
   "note": "77 options exceed the API's 8-choice limit: ours read out offline (v1.3 served Q4, v1.2 BF16 research adapter), 5 option orders; v1.0 not measured"
  }
 ],
 "summary": {
  "panel_total_correct": {
   "v1.3": 2706,
   "v1.2": 2472,
   "v1.0": 2475,
   "Jev 1.13": 2679,
   "Clef": 2704,
   "Clef-Flash": 2734,
   "Laya": 1967
  },
  "panel_total_n": 2977,
  "panel_ece10": {
   "v1.3": 0.0191,
   "v1.2": 0.0576,
   "v1.0": 0.2338
  },
  "wins_v13_vs_v10": 12,
  "rows_with_v10": 12
 }
}