{
 "validation_E_single_llm_all_anthropic_mid": {
  "arm": "E_single_llm",
  "committee": "all_anthropic",
  "n": 40,
  "scales": {
   "scale_90": 1.4900272103867178,
   "scale_50": 1.2611647423036356,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_E_single_llm_all_openai_mid": {
  "arm": "E_single_llm",
  "committee": "all_openai",
  "n": 40,
  "scales": {
   "scale_90": 1.8376879008555251,
   "scale_50": 1.471356381563702,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_E_single_llm_mixed_random_mid": {
  "arm": "E_single_llm",
  "committee": "mixed_random",
  "n": 40,
  "scales": {
   "scale_90": 1.6294725791155371,
   "scale_50": 1.2643313611844822,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_F_independent_all_anthropic_mid": {
  "arm": "F_independent",
  "committee": "all_anthropic",
  "n": 40,
  "scales": {
   "scale_90": 1.3966565212756414,
   "scale_50": 1.2851483026993815,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_F_independent_all_openai_mid": {
  "arm": "F_independent",
  "committee": "all_openai",
  "n": 40,
  "scales": {
   "scale_90": 1.7701450122812816,
   "scale_50": 1.346442789102687,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_F_independent_mixed_random_mid": {
  "arm": "F_independent",
  "committee": "mixed_random",
  "n": 40,
  "scales": {
   "scale_90": 1.584302968391311,
   "scale_50": 1.2611647423036356,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_G_debate_r1_mixed_random_mid": {
  "arm": "G_debate_r1",
  "committee": "mixed_random",
  "n": 40,
  "scales": {
   "scale_90": 1.4953562448191138,
   "scale_50": 1.2613728762745116,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_G_debate_r2_all_anthropic_mid": {
  "arm": "G_debate_r2",
  "committee": "all_anthropic",
  "n": 40,
  "scales": {
   "scale_90": 1.497771158628971,
   "scale_50": 1.3054434024714736,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_G_debate_r2_all_openai_mid": {
  "arm": "G_debate_r2",
  "committee": "all_openai",
  "n": 40,
  "scales": {
   "scale_90": 1.6689680709853856,
   "scale_50": 1.3139685740963816,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_G_debate_r2_mixed_random_mid": {
  "arm": "G_debate_r2",
  "committee": "mixed_random",
  "n": 40,
  "scales": {
   "scale_90": 1.5481787277678398,
   "scale_50": 1.2644742287419792,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_G_debate_r4_mixed_random_mid": {
  "arm": "G_debate_r4",
  "committee": "mixed_random",
  "n": 20,
  "scales": {
   "scale_90": 1.409406786569687,
   "scale_50": 1.0957996076104384,
   "n_valid": 20,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.95,
   "achieved_valid_50": 0.55,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 },
 "validation_J_cost_matched_mixed_random_mid": {
  "arm": "J_cost_matched",
  "committee": "mixed_random",
  "n": 40,
  "scales": {
   "scale_90": 1.5323539200936531,
   "scale_50": 1.2432409161671234,
   "n_valid": 40,
   "level_90": 0.9,
   "level_50": 0.5,
   "achieved_valid_90": 0.925,
   "achieved_valid_50": 0.525,
   "note": "split conformal, asymmetric scaled residual score, centre = log q50"
  }
 }
}