{
  "experiment": "expG_decision_stability",
  "author": "Simon-Pierre Boucher",
  "contact": "contact@spboucher.ai",
  "manifest": {
    "author": "Simon-Pierre Boucher",
    "contact": "contact@spboucher.ai",
    "project": "localvm-research",
    "collected_utc": "2026-08-12T04:27:44.503861+00:00",
    "chip": {
      "brand": "Apple M5 Max",
      "arch": "arm64",
      "cores_total": 18,
      "cores_performance": 6,
      "cores_efficiency": 12,
      "gpu_cores": 40
    },
    "memory": {
      "unified_bytes": 51539607552,
      "unified_gb": 48.0,
      "pagesize": 16384
    },
    "ssd": {
      "model": "APPLE SSD AP2048Z",
      "size": "2 TB",
      "smart_status": "Verified"
    },
    "os": {
      "product": "macOS",
      "version": "27.0",
      "build": "26A5388g",
      "kernel": "27.0.0"
    },
    "software": {
      "python": "3.14.4",
      "mlx": "0.32.0",
      "mlx_lm": "0.31.3",
      "torch": null,
      "numpy": "2.5.2"
    },
    "git": {
      "commit": "2b7d23b36999ff09c3e6ac03a730575e71e79650",
      "dirty_tree": true
    },
    "thermal_level_at_collect": null
  },
  "config": {
    "model": "mlx-community/Qwen3-1.7B-bf16",
    "gen_tokens": 128,
    "bits": "2,3,4,8",
    "per_domain": 8,
    "group_size": 64
  },
  "n_trajectories": 48,
  "results_by_bits": {
    "2": {
      "bits": 2,
      "n_positions": 6144,
      "agreement_rate": 0.029622395833333332,
      "mean_kl_ref_q": 11.892386436462402,
      "auroc_margin_predicts_disagreement": 0.6048301329666643,
      "median_margin_agree": 1.5,
      "median_margin_disagree": 1.0,
      "escalation_curve": [
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.9703776041666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.9703776041666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.9703776041666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.9703776041666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.9703776041666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.9703776041666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.9703776041666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.9703776041666666
        },
        {
          "tau": 0.0625,
          "escalated_frac": 0.038411458333333336,
          "residual_disagree": 0.9327799479166666
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.043619791666666664,
          "residual_disagree": 0.927734375
        },
        {
          "tau": 0.1875,
          "escalated_frac": 0.11328125,
          "residual_disagree": 0.8601888020833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.119140625,
          "residual_disagree": 0.8543294270833334
        },
        {
          "tau": 0.3125,
          "escalated_frac": 0.18912760416666666,
          "residual_disagree": 0.7853190104166666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.19352213541666666,
          "residual_disagree": 0.7809244791666666
        },
        {
          "tau": 0.4375,
          "escalated_frac": 0.25830078125,
          "residual_disagree": 0.7176106770833334
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.2630208333333333,
          "residual_disagree": 0.712890625
        },
        {
          "tau": 0.5625,
          "escalated_frac": 0.3224283854166667,
          "residual_disagree": 0.65478515625
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.328125,
          "residual_disagree": 0.6490885416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.5944010416666666
        },
        {
          "tau": 0.8125,
          "escalated_frac": 0.4353841145833333,
          "residual_disagree": 0.5445963541666666
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.4383138020833333,
          "residual_disagree": 0.5416666666666666
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.4383138020833333,
          "residual_disagree": 0.5416666666666666
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.4383138020833333,
          "residual_disagree": 0.5416666666666666
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.4383138020833333,
          "residual_disagree": 0.5416666666666666
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.4383138020833333,
          "residual_disagree": 0.5416666666666666
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.4383138020833333,
          "residual_disagree": 0.5416666666666666
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.4383138020833333,
          "residual_disagree": 0.5416666666666666
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.4383138020833333,
          "residual_disagree": 0.5416666666666666
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.4383138020833333,
          "residual_disagree": 0.5416666666666666
        },
        {
          "tau": 0.9579145728643255,
          "escalated_frac": 0.4874674479166667,
          "residual_disagree": 0.4931640625
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.4874674479166667,
          "residual_disagree": 0.4931640625
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.4874674479166667,
          "residual_disagree": 0.4931640625
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.4874674479166667,
          "residual_disagree": 0.4931640625
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.4874674479166667,
          "residual_disagree": 0.4931640625
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.4874674479166667,
          "residual_disagree": 0.4931640625
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.4874674479166667,
          "residual_disagree": 0.4931640625
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.4874674479166667,
          "residual_disagree": 0.4931640625
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.4874674479166667,
          "residual_disagree": 0.4931640625
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.5319010416666666,
          "residual_disagree": 0.4495442708333333
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.5319010416666666,
          "residual_disagree": 0.4495442708333333
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.5319010416666666,
          "residual_disagree": 0.4495442708333333
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.5319010416666666,
          "residual_disagree": 0.4495442708333333
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.5319010416666666,
          "residual_disagree": 0.4495442708333333
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.5319010416666666,
          "residual_disagree": 0.4495442708333333
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.5319010416666666,
          "residual_disagree": 0.4495442708333333
        },
        {
          "tau": 1.1875,
          "escalated_frac": 0.56689453125,
          "residual_disagree": 0.41552734375
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.5686848958333334,
          "residual_disagree": 0.4137369791666667
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.5686848958333334,
          "residual_disagree": 0.4137369791666667
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.5686848958333334,
          "residual_disagree": 0.4137369791666667
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.5686848958333334,
          "residual_disagree": 0.4137369791666667
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.5686848958333334,
          "residual_disagree": 0.4137369791666667
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.5686848958333334,
          "residual_disagree": 0.4137369791666667
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.6022135416666666,
          "residual_disagree": 0.3810221354166667
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.6022135416666666,
          "residual_disagree": 0.3810221354166667
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.6022135416666666,
          "residual_disagree": 0.3810221354166667
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.6022135416666666,
          "residual_disagree": 0.3810221354166667
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.6022135416666666,
          "residual_disagree": 0.3810221354166667
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.6022135416666666,
          "residual_disagree": 0.3810221354166667
        },
        {
          "tau": 1.4086055276381728,
          "escalated_frac": 0.6331380208333334,
          "residual_disagree": 0.3518880208333333
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.6350911458333334,
          "residual_disagree": 0.3499348958333333
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.6350911458333334,
          "residual_disagree": 0.3499348958333333
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.6350911458333334,
          "residual_disagree": 0.3499348958333333
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.6350911458333334,
          "residual_disagree": 0.3499348958333333
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.6350911458333334,
          "residual_disagree": 0.3499348958333333
        },
        {
          "tau": 1.5625,
          "escalated_frac": 0.6629231770833334,
          "residual_disagree": 0.3229166666666667
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.6640625,
          "residual_disagree": 0.32177734375
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.6640625,
          "residual_disagree": 0.32177734375
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.6640625,
          "residual_disagree": 0.32177734375
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.6640625,
          "residual_disagree": 0.32177734375
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.6640625,
          "residual_disagree": 0.32177734375
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.6640625,
          "residual_disagree": 0.32177734375
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.697265625,
          "residual_disagree": 0.2898763020833333
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.697265625,
          "residual_disagree": 0.2898763020833333
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.697265625,
          "residual_disagree": 0.2898763020833333
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.697265625,
          "residual_disagree": 0.2898763020833333
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.697265625,
          "residual_disagree": 0.2898763020833333
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.7234700520833334,
          "residual_disagree": 0.2646484375
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.7234700520833334,
          "residual_disagree": 0.2646484375
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.7234700520833334,
          "residual_disagree": 0.2646484375
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.7234700520833334,
          "residual_disagree": 0.2646484375
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.7234700520833334,
          "residual_disagree": 0.2646484375
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.7451171875,
          "residual_disagree": 0.24348958333333334
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.7451171875,
          "residual_disagree": 0.24348958333333334
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.7451171875,
          "residual_disagree": 0.24348958333333334
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.7451171875,
          "residual_disagree": 0.24348958333333334
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.76708984375,
          "residual_disagree": 0.22184244791666666
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.76708984375,
          "residual_disagree": 0.22184244791666666
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.76708984375,
          "residual_disagree": 0.22184244791666666
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.76708984375,
          "residual_disagree": 0.22184244791666666
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.787109375,
          "residual_disagree": 0.20263671875
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.787109375,
          "residual_disagree": 0.20263671875
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.787109375,
          "residual_disagree": 0.20263671875
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.8030598958333334,
          "residual_disagree": 0.18701171875
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.8030598958333334,
          "residual_disagree": 0.18701171875
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.8030598958333334,
          "residual_disagree": 0.18701171875
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.8030598958333334,
          "residual_disagree": 0.18701171875
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.8204752604166666,
          "residual_disagree": 0.169921875
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.8204752604166666,
          "residual_disagree": 0.169921875
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.8204752604166666,
          "residual_disagree": 0.169921875
        },
        {
          "tau": 2.625,
          "escalated_frac": 0.8385416666666666,
          "residual_disagree": 0.15234375
        },
        {
          "tau": 2.625,
          "escalated_frac": 0.8385416666666666,
          "residual_disagree": 0.15234375
        },
        {
          "tau": 2.625,
          "escalated_frac": 0.8385416666666666,
          "residual_disagree": 0.15234375
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.853515625,
          "residual_disagree": 0.13736979166666666
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.853515625,
          "residual_disagree": 0.13736979166666666
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.853515625,
          "residual_disagree": 0.13736979166666666
        },
        {
          "tau": 2.875,
          "escalated_frac": 0.86669921875,
          "residual_disagree": 0.12483723958333333
        },
        {
          "tau": 2.875,
          "escalated_frac": 0.86669921875,
          "residual_disagree": 0.12483723958333333
        },
        {
          "tau": 2.875,
          "escalated_frac": 0.86669921875,
          "residual_disagree": 0.12483723958333333
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.88134765625,
          "residual_disagree": 0.11051432291666667
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.88134765625,
          "residual_disagree": 0.11051432291666667
        },
        {
          "tau": 3.125,
          "escalated_frac": 0.8942057291666666,
          "residual_disagree": 0.09814453125
        },
        {
          "tau": 3.125,
          "escalated_frac": 0.8942057291666666,
          "residual_disagree": 0.09814453125
        },
        {
          "tau": 3.217650753768851,
          "escalated_frac": 0.9044596354166666,
          "residual_disagree": 0.08837890625
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.9044596354166666,
          "residual_disagree": 0.08837890625
        },
        {
          "tau": 3.375,
          "escalated_frac": 0.9130859375,
          "residual_disagree": 0.08024088541666667
        },
        {
          "tau": 3.375,
          "escalated_frac": 0.9130859375,
          "residual_disagree": 0.08024088541666667
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.92041015625,
          "residual_disagree": 0.07307942708333333
        },
        {
          "tau": 3.625,
          "escalated_frac": 0.9267578125,
          "residual_disagree": 0.06722005208333333
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.93310546875,
          "residual_disagree": 0.0615234375
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.93310546875,
          "residual_disagree": 0.0615234375
        },
        {
          "tau": 3.875,
          "escalated_frac": 0.9401041666666666,
          "residual_disagree": 0.054850260416666664
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.9456380208333334,
          "residual_disagree": 0.049641927083333336
        },
        {
          "tau": 4.125,
          "escalated_frac": 0.95166015625,
          "residual_disagree": 0.0439453125
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.9560546875,
          "residual_disagree": 0.039876302083333336
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.9637044270833334,
          "residual_disagree": 0.032552083333333336
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.9690755208333334,
          "residual_disagree": 0.02734375
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.974609375,
          "residual_disagree": 0.022298177083333332
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.97802734375,
          "residual_disagree": 0.01904296875
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.9847005208333334,
          "residual_disagree": 0.01318359375
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.98974609375,
          "residual_disagree": 0.008463541666666666
        },
        {
          "tau": 7.25,
          "escalated_frac": 0.9947916666666666,
          "residual_disagree": 0.004069010416666667
        },
        {
          "tau": 14.875,
          "escalated_frac": 0.9998372395833334,
          "residual_disagree": 0.00016276041666666666
        }
      ],
      "per_domain": {
        "code": {
          "agreement_rate": 0.017578125,
          "auroc": 0.5606914071128782
        },
        "math": {
          "agreement_rate": 0.0546875,
          "auroc": 0.6518410566706021
        },
        "chat": {
          "agreement_rate": 0.0205078125,
          "auroc": 0.5552153064615677
        },
        "french": {
          "agreement_rate": 0.0263671875,
          "auroc": 0.6100338051190609
        },
        "english": {
          "agreement_rate": 0.029296875,
          "auroc": 0.6428571428571429
        },
        "reasoning": {
          "agreement_rate": 0.029296875,
          "auroc": 0.5358484238765929
        }
      },
      "escalation_frac_for_99pct": 0.98974609375
    },
    "3": {
      "bits": 3,
      "n_positions": 6144,
      "agreement_rate": 0.73095703125,
      "mean_kl_ref_q": 0.7811363339424133,
      "auroc_margin_predicts_disagreement": 0.8521382349292252,
      "median_margin_agree": 5.25,
      "median_margin_disagree": 1.125,
      "escalation_curve": [
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.26904296875
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.26904296875
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.26904296875
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.26904296875
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.26904296875
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.020833333333333332,
          "residual_disagree": 0.2550455729166667
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.020833333333333332,
          "residual_disagree": 0.2550455729166667
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.020833333333333332,
          "residual_disagree": 0.2550455729166667
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.038248697916666664,
          "residual_disagree": 0.2421875
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.038248697916666664,
          "residual_disagree": 0.2421875
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.038248697916666664,
          "residual_disagree": 0.2421875
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.038248697916666664,
          "residual_disagree": 0.2421875
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.038248697916666664,
          "residual_disagree": 0.2421875
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.038248697916666664,
          "residual_disagree": 0.2421875
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.038248697916666664,
          "residual_disagree": 0.2421875
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.07096354166666667,
          "residual_disagree": 0.21988932291666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.07096354166666667,
          "residual_disagree": 0.21988932291666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.07096354166666667,
          "residual_disagree": 0.21988932291666666
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.08723958333333333,
          "residual_disagree": 0.20930989583333334
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.08723958333333333,
          "residual_disagree": 0.20930989583333334
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.08723958333333333,
          "residual_disagree": 0.20930989583333334
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.08723958333333333,
          "residual_disagree": 0.20930989583333334
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.08723958333333333,
          "residual_disagree": 0.20930989583333334
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.08723958333333333,
          "residual_disagree": 0.20930989583333334
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.11946614583333333,
          "residual_disagree": 0.1904296875
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.11946614583333333,
          "residual_disagree": 0.1904296875
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.11946614583333333,
          "residual_disagree": 0.1904296875
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.11946614583333333,
          "residual_disagree": 0.1904296875
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.13688151041666666,
          "residual_disagree": 0.18001302083333334
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.13688151041666666,
          "residual_disagree": 0.18001302083333334
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.13688151041666666,
          "residual_disagree": 0.18001302083333334
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.13688151041666666,
          "residual_disagree": 0.18001302083333334
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.13688151041666666,
          "residual_disagree": 0.18001302083333334
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.16552734375,
          "residual_disagree": 0.16162109375
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.16552734375,
          "residual_disagree": 0.16162109375
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.16552734375,
          "residual_disagree": 0.16162109375
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.16552734375,
          "residual_disagree": 0.16162109375
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.18131510416666666,
          "residual_disagree": 0.15185546875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.18131510416666666,
          "residual_disagree": 0.15185546875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.18131510416666666,
          "residual_disagree": 0.15185546875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.18131510416666666,
          "residual_disagree": 0.15185546875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.18131510416666666,
          "residual_disagree": 0.15185546875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.18131510416666666,
          "residual_disagree": 0.15185546875
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.21223958333333334,
          "residual_disagree": 0.13525390625
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.21223958333333334,
          "residual_disagree": 0.13525390625
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.22574869791666666,
          "residual_disagree": 0.12890625
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.22574869791666666,
          "residual_disagree": 0.12890625
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.22574869791666666,
          "residual_disagree": 0.12890625
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.22574869791666666,
          "residual_disagree": 0.12890625
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.22574869791666666,
          "residual_disagree": 0.12890625
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.22574869791666666,
          "residual_disagree": 0.12890625
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.2521158854166667,
          "residual_disagree": 0.11555989583333333
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.2521158854166667,
          "residual_disagree": 0.11555989583333333
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.2635091145833333,
          "residual_disagree": 0.10807291666666667
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.2635091145833333,
          "residual_disagree": 0.10807291666666667
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.2635091145833333,
          "residual_disagree": 0.10807291666666667
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.2635091145833333,
          "residual_disagree": 0.10807291666666667
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.28466796875,
          "residual_disagree": 0.09733072916666667
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.28466796875,
          "residual_disagree": 0.09733072916666667
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.28466796875,
          "residual_disagree": 0.09733072916666667
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.2985026041666667,
          "residual_disagree": 0.09000651041666667
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.2985026041666667,
          "residual_disagree": 0.09000651041666667
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.2985026041666667,
          "residual_disagree": 0.09000651041666667
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.2985026041666667,
          "residual_disagree": 0.09000651041666667
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.2985026041666667,
          "residual_disagree": 0.09000651041666667
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.32177734375,
          "residual_disagree": 0.07926432291666667
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.32177734375,
          "residual_disagree": 0.07926432291666667
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.33203125,
          "residual_disagree": 0.0751953125
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.33203125,
          "residual_disagree": 0.0751953125
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.33203125,
          "residual_disagree": 0.0751953125
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.33203125,
          "residual_disagree": 0.0751953125
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.3527018229166667,
          "residual_disagree": 0.068359375
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.3527018229166667,
          "residual_disagree": 0.068359375
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.3619791666666667,
          "residual_disagree": 0.06461588541666667
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.3619791666666667,
          "residual_disagree": 0.06461588541666667
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.3619791666666667,
          "residual_disagree": 0.06461588541666667
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.3811848958333333,
          "residual_disagree": 0.057942708333333336
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.3811848958333333,
          "residual_disagree": 0.057942708333333336
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.3907877604166667,
          "residual_disagree": 0.0546875
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.3907877604166667,
          "residual_disagree": 0.0546875
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.3907877604166667,
          "residual_disagree": 0.0546875
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.3907877604166667,
          "residual_disagree": 0.0546875
        },
        {
          "tau": 2.625,
          "escalated_frac": 0.4083658854166667,
          "residual_disagree": 0.04931640625
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.41650390625,
          "residual_disagree": 0.046549479166666664
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.41650390625,
          "residual_disagree": 0.046549479166666664
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.41650390625,
          "residual_disagree": 0.046549479166666664
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.41650390625,
          "residual_disagree": 0.046549479166666664
        },
        {
          "tau": 2.875,
          "escalated_frac": 0.4334309895833333,
          "residual_disagree": 0.040201822916666664
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.4404296875,
          "residual_disagree": 0.038248697916666664
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.4404296875,
          "residual_disagree": 0.038248697916666664
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.4404296875,
          "residual_disagree": 0.038248697916666664
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.4404296875,
          "residual_disagree": 0.038248697916666664
        },
        {
          "tau": 3.125,
          "escalated_frac": 0.4591471354166667,
          "residual_disagree": 0.034016927083333336
        },
        {
          "tau": 3.125,
          "escalated_frac": 0.4591471354166667,
          "residual_disagree": 0.034016927083333336
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.4685872395833333,
          "residual_disagree": 0.031575520833333336
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.4685872395833333,
          "residual_disagree": 0.031575520833333336
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.4685872395833333,
          "residual_disagree": 0.031575520833333336
        },
        {
          "tau": 3.375,
          "escalated_frac": 0.4855143229166667,
          "residual_disagree": 0.026529947916666668
        },
        {
          "tau": 3.375,
          "escalated_frac": 0.4855143229166667,
          "residual_disagree": 0.026529947916666668
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.4947916666666667,
          "residual_disagree": 0.024251302083333332
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.4947916666666667,
          "residual_disagree": 0.024251302083333332
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.4947916666666667,
          "residual_disagree": 0.024251302083333332
        },
        {
          "tau": 3.625,
          "escalated_frac": 0.5110677083333334,
          "residual_disagree": 0.021321614583333332
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.51708984375,
          "residual_disagree": 0.02001953125
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.51708984375,
          "residual_disagree": 0.02001953125
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.51708984375,
          "residual_disagree": 0.02001953125
        },
        {
          "tau": 3.875,
          "escalated_frac": 0.53173828125,
          "residual_disagree": 0.017415364583333332
        },
        {
          "tau": 3.875,
          "escalated_frac": 0.53173828125,
          "residual_disagree": 0.017415364583333332
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.5384114583333334,
          "residual_disagree": 0.016927083333333332
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.5384114583333334,
          "residual_disagree": 0.016927083333333332
        },
        {
          "tau": 4.125,
          "escalated_frac": 0.5511067708333334,
          "residual_disagree": 0.01416015625
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.5568033854166666,
          "residual_disagree": 0.013997395833333334
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.5568033854166666,
          "residual_disagree": 0.013997395833333334
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.5568033854166666,
          "residual_disagree": 0.013997395833333334
        },
        {
          "tau": 4.375,
          "escalated_frac": 0.5690104166666666,
          "residual_disagree": 0.01318359375
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.5745442708333334,
          "residual_disagree": 0.012858072916666666
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.5745442708333334,
          "residual_disagree": 0.012858072916666666
        },
        {
          "tau": 4.625,
          "escalated_frac": 0.5865885416666666,
          "residual_disagree": 0.011393229166666666
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.5926106770833334,
          "residual_disagree": 0.010904947916666666
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.5926106770833334,
          "residual_disagree": 0.010904947916666666
        },
        {
          "tau": 4.875,
          "escalated_frac": 0.6017252604166666,
          "residual_disagree": 0.009928385416666666
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.6067708333333334,
          "residual_disagree": 0.00927734375
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.6067708333333334,
          "residual_disagree": 0.00927734375
        },
        {
          "tau": 5.125,
          "escalated_frac": 0.61767578125,
          "residual_disagree": 0.0087890625
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.6214192708333334,
          "residual_disagree": 0.008626302083333334
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.6214192708333334,
          "residual_disagree": 0.008626302083333334
        },
        {
          "tau": 5.375,
          "escalated_frac": 0.63134765625,
          "residual_disagree": 0.00732421875
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.6359049479166666,
          "residual_disagree": 0.0068359375
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.6359049479166666,
          "residual_disagree": 0.0068359375
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.6466471354166666,
          "residual_disagree": 0.006510416666666667
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.6466471354166666,
          "residual_disagree": 0.006510416666666667
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.6573893229166666,
          "residual_disagree": 0.00634765625
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.6573893229166666,
          "residual_disagree": 0.00634765625
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.6661783854166666,
          "residual_disagree": 0.005208333333333333
        },
        {
          "tau": 6.375,
          "escalated_frac": 0.67236328125,
          "residual_disagree": 0.004720052083333333
        },
        {
          "tau": 6.5,
          "escalated_frac": 0.6748046875,
          "residual_disagree": 0.004720052083333333
        },
        {
          "tau": 6.625,
          "escalated_frac": 0.6818033854166666,
          "residual_disagree": 0.004069010416666667
        },
        {
          "tau": 6.75,
          "escalated_frac": 0.6863606770833334,
          "residual_disagree": 0.00390625
        },
        {
          "tau": 6.871231155778901,
          "escalated_frac": 0.693359375,
          "residual_disagree": 0.0035807291666666665
        },
        {
          "tau": 7.0,
          "escalated_frac": 0.6964518229166666,
          "residual_disagree": 0.0035807291666666665
        },
        {
          "tau": 7.125,
          "escalated_frac": 0.7024739583333334,
          "residual_disagree": 0.0032552083333333335
        },
        {
          "tau": 7.25,
          "escalated_frac": 0.7052408854166666,
          "residual_disagree": 0.0030924479166666665
        },
        {
          "tau": 7.5,
          "escalated_frac": 0.71337890625,
          "residual_disagree": 0.0027669270833333335
        },
        {
          "tau": 7.5,
          "escalated_frac": 0.71337890625,
          "residual_disagree": 0.0027669270833333335
        },
        {
          "tau": 7.75,
          "escalated_frac": 0.7233072916666666,
          "residual_disagree": 0.00244140625
        },
        {
          "tau": 7.75,
          "escalated_frac": 0.7233072916666666,
          "residual_disagree": 0.00244140625
        },
        {
          "tau": 7.875,
          "escalated_frac": 0.7314453125,
          "residual_disagree": 0.0022786458333333335
        },
        {
          "tau": 8.125,
          "escalated_frac": 0.7373046875,
          "residual_disagree": 0.0022786458333333335
        },
        {
          "tau": 8.25,
          "escalated_frac": 0.73974609375,
          "residual_disagree": 0.0022786458333333335
        },
        {
          "tau": 8.5,
          "escalated_frac": 0.7473958333333334,
          "residual_disagree": 0.0021158854166666665
        },
        {
          "tau": 8.75,
          "escalated_frac": 0.7535807291666666,
          "residual_disagree": 0.0017903645833333333
        },
        {
          "tau": 8.875,
          "escalated_frac": 0.7576497395833334,
          "residual_disagree": 0.0017903645833333333
        },
        {
          "tau": 9.0,
          "escalated_frac": 0.7600911458333334,
          "residual_disagree": 0.0017903645833333333
        },
        {
          "tau": 9.25,
          "escalated_frac": 0.7669270833333334,
          "residual_disagree": 0.0017903645833333333
        },
        {
          "tau": 9.5,
          "escalated_frac": 0.7736002604166666,
          "residual_disagree": 0.0017903645833333333
        },
        {
          "tau": 9.5,
          "escalated_frac": 0.7736002604166666,
          "residual_disagree": 0.0017903645833333333
        },
        {
          "tau": 9.75,
          "escalated_frac": 0.78076171875,
          "residual_disagree": 0.0016276041666666667
        },
        {
          "tau": 10.0,
          "escalated_frac": 0.7862955729166666,
          "residual_disagree": 0.00146484375
        },
        {
          "tau": 10.25,
          "escalated_frac": 0.7926432291666666,
          "residual_disagree": 0.00146484375
        },
        {
          "tau": 10.375,
          "escalated_frac": 0.7967122395833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 10.625,
          "escalated_frac": 0.8035481770833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 10.75,
          "escalated_frac": 0.8059895833333334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 10.979271356783897,
          "escalated_frac": 0.81396484375,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 11.125,
          "escalated_frac": 0.81787109375,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 11.25,
          "escalated_frac": 0.8203125,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 11.5,
          "escalated_frac": 0.8269856770833334,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 11.75,
          "escalated_frac": 0.8336588541666666,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 11.875,
          "escalated_frac": 0.8369140625,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 12.125,
          "escalated_frac": 0.8430989583333334,
          "residual_disagree": 0.0008138020833333334
        },
        {
          "tau": 12.364949748743697,
          "escalated_frac": 0.84912109375,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 12.5,
          "escalated_frac": 0.8518880208333334,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 12.75,
          "escalated_frac": 0.8583984375,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 13.0,
          "escalated_frac": 0.8640950520833334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 13.25,
          "escalated_frac": 0.8684895833333334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 13.5,
          "escalated_frac": 0.87353515625,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 13.75,
          "escalated_frac": 0.8790690104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.875,
          "escalated_frac": 0.8834635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.125,
          "escalated_frac": 0.8885091145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.25,
          "escalated_frac": 0.8917643229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.5,
          "escalated_frac": 0.8963216145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.75,
          "escalated_frac": 0.9021809895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.0,
          "escalated_frac": 0.9075520833333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.27763819095469,
          "escalated_frac": 0.91455078125,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.625,
          "escalated_frac": 0.9189453125,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.0,
          "escalated_frac": 0.9236653645833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.25,
          "escalated_frac": 0.92822265625,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.75,
          "escalated_frac": 0.9339192708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.25,
          "escalated_frac": 0.9392903645833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.625,
          "escalated_frac": 0.9440104166666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.875,
          "escalated_frac": 0.9484049479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.375,
          "escalated_frac": 0.9542643229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.875,
          "escalated_frac": 0.95947265625,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.375,
          "escalated_frac": 0.9640299479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.0,
          "escalated_frac": 0.9695638020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.375,
          "escalated_frac": 0.97412109375,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.125,
          "escalated_frac": 0.9793294270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.173994974874404,
          "escalated_frac": 0.98486328125,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.53266331658301,
          "escalated_frac": 0.9899088541666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.391331658291392,
          "escalated_frac": 0.9949544270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 32.375,
          "escalated_frac": 0.9998372395833334,
          "residual_disagree": 0.0
        }
      ],
      "per_domain": {
        "code": {
          "agreement_rate": 0.728515625,
          "auroc": 0.8375484598916041
        },
        "math": {
          "agreement_rate": 0.8193359375,
          "auroc": 0.8957188416068035
        },
        "chat": {
          "agreement_rate": 0.6669921875,
          "auroc": 0.8009257072686913
        },
        "french": {
          "agreement_rate": 0.728515625,
          "auroc": 0.8465605531660463
        },
        "english": {
          "agreement_rate": 0.6875,
          "auroc": 0.8543834339488636
        },
        "reasoning": {
          "agreement_rate": 0.7548828125,
          "auroc": 0.8626838055282106
        }
      },
      "escalation_frac_for_99pct": 0.6017252604166666
    },
    "4": {
      "bits": 4,
      "n_positions": 6144,
      "agreement_rate": 0.8740234375,
      "mean_kl_ref_q": 0.19848966598510742,
      "auroc_margin_predicts_disagreement": 0.8982447466304814,
      "median_margin_agree": 6.5,
      "median_margin_disagree": 0.75,
      "escalation_curve": [
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.1259765625
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.1259765625
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.1259765625
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.014485677083333334,
          "residual_disagree": 0.11783854166666667
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.014485677083333334,
          "residual_disagree": 0.11783854166666667
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.014485677083333334,
          "residual_disagree": 0.11783854166666667
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0263671875,
          "residual_disagree": 0.1083984375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0263671875,
          "residual_disagree": 0.1083984375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0263671875,
          "residual_disagree": 0.1083984375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0263671875,
          "residual_disagree": 0.1083984375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0263671875,
          "residual_disagree": 0.1083984375
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.053059895833333336,
          "residual_disagree": 0.09309895833333333
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.053059895833333336,
          "residual_disagree": 0.09309895833333333
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.06266276041666667,
          "residual_disagree": 0.08756510416666667
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.06266276041666667,
          "residual_disagree": 0.08756510416666667
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.06266276041666667,
          "residual_disagree": 0.08756510416666667
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.06266276041666667,
          "residual_disagree": 0.08756510416666667
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.06266276041666667,
          "residual_disagree": 0.08756510416666667
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.08821614583333333,
          "residual_disagree": 0.07682291666666667
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.08821614583333333,
          "residual_disagree": 0.07682291666666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09781901041666667,
          "residual_disagree": 0.07291666666666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09781901041666667,
          "residual_disagree": 0.07291666666666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09781901041666667,
          "residual_disagree": 0.07291666666666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09781901041666667,
          "residual_disagree": 0.07291666666666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09781901041666667,
          "residual_disagree": 0.07291666666666667
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.12337239583333333,
          "residual_disagree": 0.062174479166666664
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.12337239583333333,
          "residual_disagree": 0.062174479166666664
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.1318359375,
          "residual_disagree": 0.05859375
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.1318359375,
          "residual_disagree": 0.05859375
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.1318359375,
          "residual_disagree": 0.05859375
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.1318359375,
          "residual_disagree": 0.05859375
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.15543619791666666,
          "residual_disagree": 0.050455729166666664
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.15543619791666666,
          "residual_disagree": 0.050455729166666664
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.16292317708333334,
          "residual_disagree": 0.047526041666666664
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.16292317708333334,
          "residual_disagree": 0.047526041666666664
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.16292317708333334,
          "residual_disagree": 0.047526041666666664
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.16292317708333334,
          "residual_disagree": 0.047526041666666664
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.16292317708333334,
          "residual_disagree": 0.047526041666666664
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.18880208333333334,
          "residual_disagree": 0.040364583333333336
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.18880208333333334,
          "residual_disagree": 0.040364583333333336
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.19612630208333334,
          "residual_disagree": 0.0390625
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.19612630208333334,
          "residual_disagree": 0.0390625
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.19612630208333334,
          "residual_disagree": 0.0390625
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.19612630208333334,
          "residual_disagree": 0.0390625
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.21663411458333334,
          "residual_disagree": 0.03271484375
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.22412109375,
          "residual_disagree": 0.030436197916666668
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.22412109375,
          "residual_disagree": 0.030436197916666668
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.22412109375,
          "residual_disagree": 0.030436197916666668
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.22412109375,
          "residual_disagree": 0.030436197916666668
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.24235026041666666,
          "residual_disagree": 0.027180989583333332
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.24886067708333334,
          "residual_disagree": 0.026041666666666668
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.24886067708333334,
          "residual_disagree": 0.026041666666666668
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.24886067708333334,
          "residual_disagree": 0.026041666666666668
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.24886067708333334,
          "residual_disagree": 0.026041666666666668
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.2688802083333333,
          "residual_disagree": 0.02197265625
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.2688802083333333,
          "residual_disagree": 0.02197265625
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.2766927083333333,
          "residual_disagree": 0.020182291666666668
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.2766927083333333,
          "residual_disagree": 0.020182291666666668
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.2766927083333333,
          "residual_disagree": 0.020182291666666668
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.2939453125,
          "residual_disagree": 0.017903645833333332
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.3004557291666667,
          "residual_disagree": 0.017252604166666668
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.3004557291666667,
          "residual_disagree": 0.017252604166666668
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.3004557291666667,
          "residual_disagree": 0.017252604166666668
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.3004557291666667,
          "residual_disagree": 0.017252604166666668
        },
        {
          "tau": 2.625,
          "escalated_frac": 0.3196614583333333,
          "residual_disagree": 0.014811197916666666
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.3263346354166667,
          "residual_disagree": 0.013834635416666666
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.3263346354166667,
          "residual_disagree": 0.013834635416666666
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.3263346354166667,
          "residual_disagree": 0.013834635416666666
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.3263346354166667,
          "residual_disagree": 0.013834635416666666
        },
        {
          "tau": 2.875,
          "escalated_frac": 0.3448893229166667,
          "residual_disagree": 0.01123046875
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3512369791666667,
          "residual_disagree": 0.01025390625
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3512369791666667,
          "residual_disagree": 0.01025390625
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3512369791666667,
          "residual_disagree": 0.01025390625
        },
        {
          "tau": 3.125,
          "escalated_frac": 0.3658854166666667,
          "residual_disagree": 0.009114583333333334
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.37060546875,
          "residual_disagree": 0.00830078125
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.37060546875,
          "residual_disagree": 0.00830078125
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.37060546875,
          "residual_disagree": 0.00830078125
        },
        {
          "tau": 3.375,
          "escalated_frac": 0.38623046875,
          "residual_disagree": 0.007161458333333333
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.3914388020833333,
          "residual_disagree": 0.006673177083333333
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.3914388020833333,
          "residual_disagree": 0.006673177083333333
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.3914388020833333,
          "residual_disagree": 0.006673177083333333
        },
        {
          "tau": 3.5521356783919487,
          "escalated_frac": 0.4070638020833333,
          "residual_disagree": 0.00537109375
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.4112955729166667,
          "residual_disagree": 0.005045572916666667
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.4112955729166667,
          "residual_disagree": 0.005045572916666667
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.4112955729166667,
          "residual_disagree": 0.005045572916666667
        },
        {
          "tau": 3.875,
          "escalated_frac": 0.4231770833333333,
          "residual_disagree": 0.004231770833333333
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.42724609375,
          "residual_disagree": 0.004231770833333333
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.42724609375,
          "residual_disagree": 0.004231770833333333
        },
        {
          "tau": 4.125,
          "escalated_frac": 0.4386393229166667,
          "residual_disagree": 0.0037434895833333335
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.44287109375,
          "residual_disagree": 0.0037434895833333335
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.44287109375,
          "residual_disagree": 0.0037434895833333335
        },
        {
          "tau": 4.375,
          "escalated_frac": 0.4558919270833333,
          "residual_disagree": 0.0029296875
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.4591471354166667,
          "residual_disagree": 0.0029296875
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.4591471354166667,
          "residual_disagree": 0.0029296875
        },
        {
          "tau": 4.625,
          "escalated_frac": 0.4698893229166667,
          "residual_disagree": 0.00244140625
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.4737955729166667,
          "residual_disagree": 0.00244140625
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.4737955729166667,
          "residual_disagree": 0.00244140625
        },
        {
          "tau": 4.875,
          "escalated_frac": 0.4850260416666667,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.4889322916666667,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.4889322916666667,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 5.125,
          "escalated_frac": 0.49853515625,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.50390625,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.50390625,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.5149739583333334,
          "residual_disagree": 0.0017903645833333333
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.5149739583333334,
          "residual_disagree": 0.0017903645833333333
        },
        {
          "tau": 5.625,
          "escalated_frac": 0.5255533854166666,
          "residual_disagree": 0.0016276041666666667
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.5284830729166666,
          "residual_disagree": 0.0016276041666666667
        },
        {
          "tau": 5.875,
          "escalated_frac": 0.537109375,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.5413411458333334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 6.125,
          "escalated_frac": 0.5472005208333334,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.5491536458333334,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 6.375,
          "escalated_frac": 0.55712890625,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 6.5,
          "escalated_frac": 0.5597330729166666,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 6.625,
          "escalated_frac": 0.5665690104166666,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 6.75,
          "escalated_frac": 0.57080078125,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 6.75,
          "escalated_frac": 0.57080078125,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 7.0,
          "escalated_frac": 0.58154296875,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 7.0,
          "escalated_frac": 0.58154296875,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 7.197864321608051,
          "escalated_frac": 0.5929361979166666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 7.25,
          "escalated_frac": 0.5929361979166666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 7.375,
          "escalated_frac": 0.6013997395833334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 7.5,
          "escalated_frac": 0.6046549479166666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 7.75,
          "escalated_frac": 0.6123046875,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 7.875,
          "escalated_frac": 0.6178385416666666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 8.0,
          "escalated_frac": 0.6214192708333334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 8.125,
          "escalated_frac": 0.6272786458333334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 8.25,
          "escalated_frac": 0.6316731770833334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 8.375,
          "escalated_frac": 0.6376953125,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 8.5,
          "escalated_frac": 0.640625,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 8.625,
          "escalated_frac": 0.6466471354166666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 8.75,
          "escalated_frac": 0.650390625,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 8.875,
          "escalated_frac": 0.6555989583333334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 9.0,
          "escalated_frac": 0.6598307291666666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 9.125,
          "escalated_frac": 0.6653645833333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.25,
          "escalated_frac": 0.6686197916666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.5,
          "escalated_frac": 0.6770833333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.5,
          "escalated_frac": 0.6770833333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.75,
          "escalated_frac": 0.6861979166666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.9962311557789,
          "escalated_frac": 0.693359375,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.125,
          "escalated_frac": 0.6971028645833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.25,
          "escalated_frac": 0.6988932291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.5,
          "escalated_frac": 0.7068684895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.625,
          "escalated_frac": 0.7120768229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.75,
          "escalated_frac": 0.71435546875,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.0,
          "escalated_frac": 0.7220052083333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.125,
          "escalated_frac": 0.7275390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.375,
          "escalated_frac": 0.7327473958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.5,
          "escalated_frac": 0.7360026041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.75,
          "escalated_frac": 0.7410481770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.0,
          "escalated_frac": 0.7478841145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.25,
          "escalated_frac": 0.7529296875,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.375,
          "escalated_frac": 0.7566731770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.75,
          "escalated_frac": 0.7630208333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.875,
          "escalated_frac": 0.7667643229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.125,
          "escalated_frac": 0.7734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.343592964824097,
          "escalated_frac": 0.77880859375,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.5,
          "escalated_frac": 0.78173828125,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.625,
          "escalated_frac": 0.7880859375,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.875,
          "escalated_frac": 0.79345703125,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.0,
          "escalated_frac": 0.7972005208333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.25,
          "escalated_frac": 0.8025716145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.375,
          "escalated_frac": 0.8079427083333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.5,
          "escalated_frac": 0.8103841145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.75,
          "escalated_frac": 0.8170572916666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.875,
          "escalated_frac": 0.8212890625,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.180276381909607,
          "escalated_frac": 0.8291015625,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.5,
          "escalated_frac": 0.8331705729166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.75,
          "escalated_frac": 0.8377278645833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.0,
          "escalated_frac": 0.8435872395833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.125,
          "escalated_frac": 0.84814453125,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.375,
          "escalated_frac": 0.8528645833333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.625,
          "escalated_frac": 0.8583984375,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.815954773869294,
          "escalated_frac": 0.8642578125,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.125,
          "escalated_frac": 0.8680013020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.375,
          "escalated_frac": 0.8738606770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.625,
          "escalated_frac": 0.8782552083333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.75062814070361,
          "escalated_frac": 0.8844401041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.125,
          "escalated_frac": 0.8880208333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.5,
          "escalated_frac": 0.8937174479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.875,
          "escalated_frac": 0.8990885416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.125,
          "escalated_frac": 0.9025065104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.29396984924631,
          "escalated_frac": 0.9095052083333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.625,
          "escalated_frac": 0.9138997395833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.875,
          "escalated_frac": 0.9181315104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.25,
          "escalated_frac": 0.9235026041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.625,
          "escalated_frac": 0.9285481770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.0,
          "escalated_frac": 0.93310546875,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.375,
          "escalated_frac": 0.9383138020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.875,
          "escalated_frac": 0.9435221354166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.25,
          "escalated_frac": 0.9495442708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.625,
          "escalated_frac": 0.9542643229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.130653266331706,
          "escalated_frac": 0.9597981770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.75,
          "escalated_frac": 0.9645182291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.375,
          "escalated_frac": 0.9689127604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.206658291457302,
          "escalated_frac": 0.9747721354166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.875,
          "escalated_frac": 0.97900390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 26.875,
          "escalated_frac": 0.9845377604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 28.125,
          "escalated_frac": 0.9894205729166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 29.766331658291392,
          "escalated_frac": 0.9949544270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 35.25,
          "escalated_frac": 0.9998372395833334,
          "residual_disagree": 0.0
        }
      ],
      "per_domain": {
        "code": {
          "agreement_rate": 0.857421875,
          "auroc": 0.8844041563952945
        },
        "math": {
          "agreement_rate": 0.9365234375,
          "auroc": 0.9211117349803482
        },
        "chat": {
          "agreement_rate": 0.8271484375,
          "auroc": 0.880055229824105
        },
        "french": {
          "agreement_rate": 0.87109375,
          "auroc": 0.9024663677130045
        },
        "english": {
          "agreement_rate": 0.84765625,
          "auroc": 0.9031445704832801
        },
        "reasoning": {
          "agreement_rate": 0.904296875,
          "auroc": 0.904030942830696
        }
      },
      "escalation_frac_for_99pct": 0.3658854166666667
    },
    "8": {
      "bits": 8,
      "n_positions": 6144,
      "agreement_rate": 0.9803059895833334,
      "mean_kl_ref_q": 0.003948633559048176,
      "auroc_margin_predicts_disagreement": 0.9745905159697743,
      "median_margin_agree": 6.25,
      "median_margin_disagree": 0.25,
      "escalation_curve": [
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.019694010416666668
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.019694010416666668
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.019694010416666668
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.012858072916666666,
          "residual_disagree": 0.014973958333333334
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.012858072916666666,
          "residual_disagree": 0.014973958333333334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.022623697916666668,
          "residual_disagree": 0.011393229166666666
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.022623697916666668,
          "residual_disagree": 0.011393229166666666
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.022623697916666668,
          "residual_disagree": 0.011393229166666666
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.022623697916666668,
          "residual_disagree": 0.011393229166666666
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.022623697916666668,
          "residual_disagree": 0.011393229166666666
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.022623697916666668,
          "residual_disagree": 0.011393229166666666
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.051106770833333336,
          "residual_disagree": 0.0037434895833333335
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059895833333333336,
          "residual_disagree": 0.0027669270833333335
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059895833333333336,
          "residual_disagree": 0.0027669270833333335
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059895833333333336,
          "residual_disagree": 0.0027669270833333335
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059895833333333336,
          "residual_disagree": 0.0027669270833333335
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059895833333333336,
          "residual_disagree": 0.0027669270833333335
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.08154296875,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08935546875,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08935546875,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08935546875,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08935546875,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08935546875,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.1142578125,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.1142578125,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12109375,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12109375,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12109375,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12109375,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.14111328125,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.14713541666666666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.14713541666666666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.14713541666666666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.14713541666666666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.17041015625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.1748046875,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.1748046875,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.1748046875,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.1748046875,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.19547526041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.20084635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.20084635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.20084635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.20084635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.20084635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.22233072916666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.22672526041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.22672526041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.22672526041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.2431640625,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.24755859375,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.24755859375,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.24755859375,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.2644856770833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2696940104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2696940104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2696940104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2696940104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.625,
          "escalated_frac": 0.2872721354166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.2931315104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.2931315104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.2931315104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.875,
          "escalated_frac": 0.3089192708333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3136393229166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3136393229166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3136393229166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.125,
          "escalated_frac": 0.3304036458333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.33447265625,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.33447265625,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.33447265625,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.375,
          "escalated_frac": 0.3504231770833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.35400390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.35400390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.35400390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.37158203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.37158203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.37158203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.37158203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.875,
          "escalated_frac": 0.3888346354166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.3929036458333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.3929036458333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.177135678391949,
          "escalated_frac": 0.4070638020833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.4070638020833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.4070638020833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.375,
          "escalated_frac": 0.42138671875,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.42529296875,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.42529296875,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.625,
          "escalated_frac": 0.4358723958333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.43896484375,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.43896484375,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.875,
          "escalated_frac": 0.4503580729166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.4529622395833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.4529622395833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.4658203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.4658203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.375,
          "escalated_frac": 0.4752604166666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.4775390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.625,
          "escalated_frac": 0.4871419270833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.4900716145833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.875,
          "escalated_frac": 0.4972330729166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.49951171875,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.125,
          "escalated_frac": 0.5066731770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.50927734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.50927734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.5,
          "escalated_frac": 0.52099609375,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.5,
          "escalated_frac": 0.52099609375,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.75,
          "escalated_frac": 0.53173828125,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.75,
          "escalated_frac": 0.53173828125,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.875,
          "escalated_frac": 0.541015625,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.0,
          "escalated_frac": 0.54296875,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.125,
          "escalated_frac": 0.5499674479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.25,
          "escalated_frac": 0.5540364583333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.5,
          "escalated_frac": 0.5626627604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.5,
          "escalated_frac": 0.5626627604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.75,
          "escalated_frac": 0.5709635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.75,
          "escalated_frac": 0.5709635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.0,
          "escalated_frac": 0.5802408854166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.125,
          "escalated_frac": 0.5852864583333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.25,
          "escalated_frac": 0.5880533854166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.5,
          "escalated_frac": 0.59765625,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.5,
          "escalated_frac": 0.59765625,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.648869346733647,
          "escalated_frac": 0.6080729166666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.75,
          "escalated_frac": 0.6080729166666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.0,
          "escalated_frac": 0.6160481770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.125,
          "escalated_frac": 0.6222330729166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.25,
          "escalated_frac": 0.6253255208333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.375,
          "escalated_frac": 0.6326497395833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.5,
          "escalated_frac": 0.63623046875,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.625,
          "escalated_frac": 0.6416015625,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.75,
          "escalated_frac": 0.6433919270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.0,
          "escalated_frac": 0.6531575520833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.0,
          "escalated_frac": 0.6531575520833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.25,
          "escalated_frac": 0.66064453125,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.5,
          "escalated_frac": 0.6673177083333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.75,
          "escalated_frac": 0.6726888020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.75,
          "escalated_frac": 0.6726888020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.0,
          "escalated_frac": 0.681640625,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.25,
          "escalated_frac": 0.6883138020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.3712311557789,
          "escalated_frac": 0.693359375,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.5,
          "escalated_frac": 0.6954752604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.75,
          "escalated_frac": 0.7021484375,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.875,
          "escalated_frac": 0.70556640625,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.0,
          "escalated_frac": 0.708984375,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.25,
          "escalated_frac": 0.71728515625,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.375,
          "escalated_frac": 0.7205403645833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.625,
          "escalated_frac": 0.7278645833333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.75,
          "escalated_frac": 0.7301432291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.0,
          "escalated_frac": 0.7371419270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.25,
          "escalated_frac": 0.7425130208333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.5,
          "escalated_frac": 0.7478841145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.75,
          "escalated_frac": 0.7527669270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.0,
          "escalated_frac": 0.7584635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.125,
          "escalated_frac": 0.7633463541666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.25,
          "escalated_frac": 0.76611328125,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.5,
          "escalated_frac": 0.7716471354166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.75,
          "escalated_frac": 0.7786458333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.875,
          "escalated_frac": 0.7830403645833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.0,
          "escalated_frac": 0.7867838541666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.25,
          "escalated_frac": 0.7937825520833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.375,
          "escalated_frac": 0.796875,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.5,
          "escalated_frac": 0.8009440104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.75,
          "escalated_frac": 0.8082682291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.875,
          "escalated_frac": 0.8136393229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.087939698492505,
          "escalated_frac": 0.8190104166666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.375,
          "escalated_frac": 0.8230794270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.625,
          "escalated_frac": 0.828125,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.75,
          "escalated_frac": 0.8312174479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.0,
          "escalated_frac": 0.8380533854166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.25,
          "escalated_frac": 0.8435872395833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.5,
          "escalated_frac": 0.8479817708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.75,
          "escalated_frac": 0.8528645833333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.0,
          "escalated_frac": 0.8575846354166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.190954773869294,
          "escalated_frac": 0.8642578125,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.5,
          "escalated_frac": 0.86767578125,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.75,
          "escalated_frac": 0.8733723958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.125,
          "escalated_frac": 0.8790690104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.5,
          "escalated_frac": 0.88427734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.75,
          "escalated_frac": 0.8885091145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.125,
          "escalated_frac": 0.8935546875,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.5,
          "escalated_frac": 0.8982747395833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.8103015075377,
          "escalated_frac": 0.9044596354166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.125,
          "escalated_frac": 0.9090169270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.5,
          "escalated_frac": 0.91357421875,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.875,
          "escalated_frac": 0.9182942708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.25,
          "escalated_frac": 0.9225260416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.625,
          "escalated_frac": 0.9285481770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.0,
          "escalated_frac": 0.93310546875,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.5,
          "escalated_frac": 0.9391276041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.875,
          "escalated_frac": 0.9436848958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.375,
          "escalated_frac": 0.9495442708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.88599246231155,
          "escalated_frac": 0.9547526041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.25,
          "escalated_frac": 0.9578450520833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.875,
          "escalated_frac": 0.9645182291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 26.695979899497388,
          "escalated_frac": 0.9697265625,
          "residual_disagree": 0.0
        },
        {
          "tau": 27.25,
          "escalated_frac": 0.9744466145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 28.25,
          "escalated_frac": 0.9794921875,
          "residual_disagree": 0.0
        },
        {
          "tau": 29.5,
          "escalated_frac": 0.9845377604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 30.5,
          "escalated_frac": 0.98974609375,
          "residual_disagree": 0.0
        },
        {
          "tau": 32.032663316582784,
          "escalated_frac": 0.9949544270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 37.5,
          "escalated_frac": 0.9998372395833334,
          "residual_disagree": 0.0
        }
      ],
      "per_domain": {
        "code": {
          "agreement_rate": 0.982421875,
          "auroc": 0.9611221559531699
        },
        "math": {
          "agreement_rate": 0.9912109375,
          "auroc": 0.9871373836891079
        },
        "chat": {
          "agreement_rate": 0.974609375,
          "auroc": 0.968937875751503
        },
        "french": {
          "agreement_rate": 0.978515625,
          "auroc": 0.981854472872437
        },
        "english": {
          "agreement_rate": 0.9677734375,
          "auroc": 0.9724490107941167
        },
        "reasoning": {
          "agreement_rate": 0.9873046875,
          "auroc": 0.9800654340713688
        }
      },
      "escalation_frac_for_99pct": 0.051106770833333336
    }
  }
}