{
  "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": "2026-08-12T05:11:29.643082+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": "74574f3169ebc78c217e870cbb54b9696bfa0ac7",
      "dirty_tree": true
    },
    "thermal_level_at_collect": null
  },
  "config": {
    "model": "mlx-community/Qwen3-8B-bf16",
    "gen_tokens": 128,
    "bits": "3,4,8",
    "per_domain": 8,
    "group_size": 64
  },
  "n_trajectories": 48,
  "results_by_bits": {
    "3": {
      "bits": 3,
      "n_positions": 6144,
      "agreement_rate": 0.8527018229166666,
      "mean_kl_ref_q": 0.34334734082221985,
      "auroc_margin_predicts_disagreement": 0.883382071775749,
      "median_margin_agree": 7.5,
      "median_margin_disagree": 1.0,
      "escalation_curve": [
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.14729817708333334
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.14729817708333334
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.14729817708333334
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.01318359375,
          "residual_disagree": 0.13883463541666666
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.01318359375,
          "residual_disagree": 0.13883463541666666
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.021484375,
          "residual_disagree": 0.13264973958333334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.021484375,
          "residual_disagree": 0.13264973958333334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.021484375,
          "residual_disagree": 0.13264973958333334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.021484375,
          "residual_disagree": 0.13264973958333334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.021484375,
          "residual_disagree": 0.13264973958333334
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.049967447916666664,
          "residual_disagree": 0.11865234375
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.049967447916666664,
          "residual_disagree": 0.11865234375
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.0576171875,
          "residual_disagree": 0.11474609375
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.0576171875,
          "residual_disagree": 0.11474609375
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.0576171875,
          "residual_disagree": 0.11474609375
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.0576171875,
          "residual_disagree": 0.11474609375
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.0576171875,
          "residual_disagree": 0.11474609375
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.0576171875,
          "residual_disagree": 0.11474609375
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.08821614583333333,
          "residual_disagree": 0.09977213541666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.0927734375,
          "residual_disagree": 0.09781901041666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.0927734375,
          "residual_disagree": 0.09781901041666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.0927734375,
          "residual_disagree": 0.09781901041666667
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.0927734375,
          "residual_disagree": 0.09781901041666667
        },
        {
          "tau": 0.8743718592964882,
          "escalated_frac": 0.11555989583333333,
          "residual_disagree": 0.08740234375
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.11555989583333333,
          "residual_disagree": 0.08740234375
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12239583333333333,
          "residual_disagree": 0.0849609375
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12239583333333333,
          "residual_disagree": 0.0849609375
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12239583333333333,
          "residual_disagree": 0.0849609375
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12239583333333333,
          "residual_disagree": 0.0849609375
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12239583333333333,
          "residual_disagree": 0.0849609375
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.14762369791666666,
          "residual_disagree": 0.0732421875
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.15348307291666666,
          "residual_disagree": 0.07063802083333333
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.15348307291666666,
          "residual_disagree": 0.07063802083333333
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.15348307291666666,
          "residual_disagree": 0.07063802083333333
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.15348307291666666,
          "residual_disagree": 0.07063802083333333
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.1748046875,
          "residual_disagree": 0.0625
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.18017578125,
          "residual_disagree": 0.060709635416666664
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.18017578125,
          "residual_disagree": 0.060709635416666664
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.18017578125,
          "residual_disagree": 0.060709635416666664
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.18017578125,
          "residual_disagree": 0.060709635416666664
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.18017578125,
          "residual_disagree": 0.060709635416666664
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.20361328125,
          "residual_disagree": 0.0517578125
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.20735677083333334,
          "residual_disagree": 0.049641927083333336
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.20735677083333334,
          "residual_disagree": 0.049641927083333336
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.20735677083333334,
          "residual_disagree": 0.049641927083333336
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.22526041666666666,
          "residual_disagree": 0.043782552083333336
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.22526041666666666,
          "residual_disagree": 0.043782552083333336
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.23177083333333334,
          "residual_disagree": 0.041666666666666664
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.23177083333333334,
          "residual_disagree": 0.041666666666666664
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.23177083333333334,
          "residual_disagree": 0.041666666666666664
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.24837239583333334,
          "residual_disagree": 0.037109375
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.2535807291666667,
          "residual_disagree": 0.035807291666666664
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.2535807291666667,
          "residual_disagree": 0.035807291666666664
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.2535807291666667,
          "residual_disagree": 0.035807291666666664
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.2706705729166667,
          "residual_disagree": 0.03076171875
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.2706705729166667,
          "residual_disagree": 0.03076171875
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2766927083333333,
          "residual_disagree": 0.028971354166666668
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2766927083333333,
          "residual_disagree": 0.028971354166666668
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2766927083333333,
          "residual_disagree": 0.028971354166666668
        },
        {
          "tau": 2.625,
          "escalated_frac": 0.2955729166666667,
          "residual_disagree": 0.02490234375
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.29931640625,
          "residual_disagree": 0.024251302083333332
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.29931640625,
          "residual_disagree": 0.024251302083333332
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.29931640625,
          "residual_disagree": 0.024251302083333332
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.29931640625,
          "residual_disagree": 0.024251302083333332
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3214518229166667,
          "residual_disagree": 0.020182291666666668
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3214518229166667,
          "residual_disagree": 0.020182291666666668
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3214518229166667,
          "residual_disagree": 0.020182291666666668
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3214518229166667,
          "residual_disagree": 0.020182291666666668
        },
        {
          "tau": 3.125,
          "escalated_frac": 0.3390299479166667,
          "residual_disagree": 0.017740885416666668
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.3426106770833333,
          "residual_disagree": 0.017252604166666668
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.3426106770833333,
          "residual_disagree": 0.017252604166666668
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.3426106770833333,
          "residual_disagree": 0.017252604166666668
        },
        {
          "tau": 3.4491206030151034,
          "escalated_frac": 0.36181640625,
          "residual_disagree": 0.01416015625
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.36181640625,
          "residual_disagree": 0.01416015625
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.36181640625,
          "residual_disagree": 0.01416015625
        },
        {
          "tau": 3.625,
          "escalated_frac": 0.3758138020833333,
          "residual_disagree": 0.012532552083333334
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.3782552083333333,
          "residual_disagree": 0.01220703125
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.3782552083333333,
          "residual_disagree": 0.01220703125
        },
        {
          "tau": 3.875,
          "escalated_frac": 0.3904622395833333,
          "residual_disagree": 0.0107421875
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.39306640625,
          "residual_disagree": 0.010416666666666666
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.39306640625,
          "residual_disagree": 0.010416666666666666
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.39306640625,
          "residual_disagree": 0.010416666666666666
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.4098307291666667,
          "residual_disagree": 0.00927734375
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.4098307291666667,
          "residual_disagree": 0.00927734375
        },
        {
          "tau": 4.375,
          "escalated_frac": 0.4195963541666667,
          "residual_disagree": 0.008138020833333334
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.4225260416666667,
          "residual_disagree": 0.007975260416666666
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.4225260416666667,
          "residual_disagree": 0.007975260416666666
        },
        {
          "tau": 4.625,
          "escalated_frac": 0.4342447916666667,
          "residual_disagree": 0.006673177083333333
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.4378255208333333,
          "residual_disagree": 0.006510416666666667
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.4378255208333333,
          "residual_disagree": 0.006510416666666667
        },
        {
          "tau": 4.875,
          "escalated_frac": 0.4493815104166667,
          "residual_disagree": 0.005533854166666667
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.45263671875,
          "residual_disagree": 0.00537109375
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.45263671875,
          "residual_disagree": 0.00537109375
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.4659830729166667,
          "residual_disagree": 0.005045572916666667
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.4659830729166667,
          "residual_disagree": 0.005045572916666667
        },
        {
          "tau": 5.375,
          "escalated_frac": 0.4757486979166667,
          "residual_disagree": 0.00390625
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.4788411458333333,
          "residual_disagree": 0.00390625
        },
        {
          "tau": 5.625,
          "escalated_frac": 0.4869791666666667,
          "residual_disagree": 0.0035807291666666665
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.4900716145833333,
          "residual_disagree": 0.00341796875
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.4900716145833333,
          "residual_disagree": 0.00341796875
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.5009765625,
          "residual_disagree": 0.00341796875
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.5009765625,
          "residual_disagree": 0.00341796875
        },
        {
          "tau": 6.125,
          "escalated_frac": 0.51171875,
          "residual_disagree": 0.0030924479166666665
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.5152994791666666,
          "residual_disagree": 0.0029296875
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.5152994791666666,
          "residual_disagree": 0.0029296875
        },
        {
          "tau": 6.5,
          "escalated_frac": 0.5242513020833334,
          "residual_disagree": 0.0026041666666666665
        },
        {
          "tau": 6.518844221105496,
          "escalated_frac": 0.53271484375,
          "residual_disagree": 0.0021158854166666665
        },
        {
          "tau": 6.75,
          "escalated_frac": 0.5349934895833334,
          "residual_disagree": 0.0021158854166666665
        },
        {
          "tau": 6.875,
          "escalated_frac": 0.5423177083333334,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 7.0,
          "escalated_frac": 0.54443359375,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 7.125,
          "escalated_frac": 0.5519205729166666,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 7.25,
          "escalated_frac": 0.5545247395833334,
          "residual_disagree": 0.001953125
        },
        {
          "tau": 7.375,
          "escalated_frac": 0.5621744791666666,
          "residual_disagree": 0.0016276041666666667
        },
        {
          "tau": 7.5,
          "escalated_frac": 0.5647786458333334,
          "residual_disagree": 0.0016276041666666667
        },
        {
          "tau": 7.513190954773904,
          "escalated_frac": 0.5729166666666666,
          "residual_disagree": 0.00146484375
        },
        {
          "tau": 7.75,
          "escalated_frac": 0.5755208333333334,
          "residual_disagree": 0.00146484375
        },
        {
          "tau": 7.875,
          "escalated_frac": 0.5817057291666666,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 8.0,
          "escalated_frac": 0.5841471354166666,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 8.125,
          "escalated_frac": 0.59130859375,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 8.25,
          "escalated_frac": 0.59423828125,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 8.375,
          "escalated_frac": 0.6013997395833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 8.5,
          "escalated_frac": 0.603515625,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 8.625,
          "escalated_frac": 0.6101888020833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 8.75,
          "escalated_frac": 0.61328125,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 8.875,
          "escalated_frac": 0.6209309895833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 9.0,
          "escalated_frac": 0.6238606770833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 9.25,
          "escalated_frac": 0.6326497395833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 9.25,
          "escalated_frac": 0.6326497395833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 9.375,
          "escalated_frac": 0.6404622395833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 9.5,
          "escalated_frac": 0.6437174479166666,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 9.75,
          "escalated_frac": 0.6502278645833334,
          "residual_disagree": 0.0013020833333333333
        },
        {
          "tau": 9.875,
          "escalated_frac": 0.6560872395833334,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 10.0,
          "escalated_frac": 0.65966796875,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 10.25,
          "escalated_frac": 0.6669921875,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 10.375,
          "escalated_frac": 0.6715494791666666,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 10.5,
          "escalated_frac": 0.6741536458333334,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 10.75,
          "escalated_frac": 0.6818033854166666,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 10.875,
          "escalated_frac": 0.6865234375,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 11.1212311557789,
          "escalated_frac": 0.693359375,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 11.25,
          "escalated_frac": 0.6941731770833334,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 11.5,
          "escalated_frac": 0.7001953125,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 11.75,
          "escalated_frac": 0.7080078125,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 11.75,
          "escalated_frac": 0.7080078125,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 12.0,
          "escalated_frac": 0.7154947916666666,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 12.25,
          "escalated_frac": 0.7220052083333334,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 12.375,
          "escalated_frac": 0.7278645833333334,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 12.5,
          "escalated_frac": 0.7296549479166666,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 12.75,
          "escalated_frac": 0.736328125,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 13.0,
          "escalated_frac": 0.7405598958333334,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 13.25,
          "escalated_frac": 0.7467447916666666,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 13.5,
          "escalated_frac": 0.7517903645833334,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 13.75,
          "escalated_frac": 0.7576497395833334,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 13.875,
          "escalated_frac": 0.76171875,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 14.001256281406995,
          "escalated_frac": 0.7688802083333334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 14.25,
          "escalated_frac": 0.7711588541666666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 14.5,
          "escalated_frac": 0.7760416666666666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 14.75,
          "escalated_frac": 0.7822265625,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 15.0,
          "escalated_frac": 0.7884114583333334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 15.125,
          "escalated_frac": 0.7921549479166666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 15.375,
          "escalated_frac": 0.7981770833333334,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 15.5,
          "escalated_frac": 0.7999674479166666,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 15.75,
          "escalated_frac": 0.8063151041666666,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 16.0,
          "escalated_frac": 0.81201171875,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.25,
          "escalated_frac": 0.8172200520833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.5,
          "escalated_frac": 0.82177734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.75,
          "escalated_frac": 0.8279622395833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.0,
          "escalated_frac": 0.8317057291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.25,
          "escalated_frac": 0.8365885416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.5,
          "escalated_frac": 0.84326171875,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.75,
          "escalated_frac": 0.8473307291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.0,
          "escalated_frac": 0.8522135416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.25,
          "escalated_frac": 0.857421875,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.5,
          "escalated_frac": 0.86279296875,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.75,
          "escalated_frac": 0.8678385416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.0,
          "escalated_frac": 0.87255859375,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.25,
          "escalated_frac": 0.8780924479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.37562814070361,
          "escalated_frac": 0.8844401041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.75,
          "escalated_frac": 0.8880208333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.9679648241206,
          "escalated_frac": 0.8943684895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.25,
          "escalated_frac": 0.8984375,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.5,
          "escalated_frac": 0.9020182291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.75,
          "escalated_frac": 0.90771484375,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.125,
          "escalated_frac": 0.9140625,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.375,
          "escalated_frac": 0.9178059895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.744974874371906,
          "escalated_frac": 0.9244791666666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.0,
          "escalated_frac": 0.9287109375,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.462311557789008,
          "escalated_frac": 0.9345703125,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.75,
          "escalated_frac": 0.93798828125,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.25,
          "escalated_frac": 0.9443359375,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.625,
          "escalated_frac": 0.9482421875,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.021984924623098,
          "escalated_frac": 0.9547526041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.5,
          "escalated_frac": 0.9581705729166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.125,
          "escalated_frac": 0.9640299479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.722989949748694,
          "escalated_frac": 0.9697265625,
          "residual_disagree": 0.0
        },
        {
          "tau": 26.375,
          "escalated_frac": 0.974609375,
          "residual_disagree": 0.0
        },
        {
          "tau": 27.125,
          "escalated_frac": 0.9794921875,
          "residual_disagree": 0.0
        },
        {
          "tau": 28.25,
          "escalated_frac": 0.984375,
          "residual_disagree": 0.0
        },
        {
          "tau": 29.28266331658301,
          "escalated_frac": 0.9899088541666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 31.25,
          "escalated_frac": 0.99462890625,
          "residual_disagree": 0.0
        },
        {
          "tau": 40.0,
          "escalated_frac": 0.9998372395833334,
          "residual_disagree": 0.0
        }
      ],
      "per_domain": {
        "code": {
          "agreement_rate": 0.8603515625,
          "auroc": 0.8824444567917894
        },
        "math": {
          "agreement_rate": 0.90234375,
          "auroc": 0.9081818181818182
        },
        "chat": {
          "agreement_rate": 0.8173828125,
          "auroc": 0.8669426714967512
        },
        "french": {
          "agreement_rate": 0.8359375,
          "auroc": 0.8821345126835781
        },
        "english": {
          "agreement_rate": 0.822265625,
          "auroc": 0.8701547858317454
        },
        "reasoning": {
          "agreement_rate": 0.8779296875,
          "auroc": 0.8931968854282536
        }
      },
      "escalation_frac_for_99pct": 0.4098307291666667
    },
    "4": {
      "bits": 4,
      "n_positions": 6144,
      "agreement_rate": 0.9212239583333334,
      "mean_kl_ref_q": 0.08975151926279068,
      "auroc_margin_predicts_disagreement": 0.9199168443185468,
      "median_margin_agree": 7.125,
      "median_margin_disagree": 0.5,
      "escalation_curve": [
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.07877604166666667
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.07877604166666667
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.07877604166666667
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.013834635416666666,
          "residual_disagree": 0.07177734375
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.013834635416666666,
          "residual_disagree": 0.07177734375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0205078125,
          "residual_disagree": 0.068359375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0205078125,
          "residual_disagree": 0.068359375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0205078125,
          "residual_disagree": 0.068359375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0205078125,
          "residual_disagree": 0.068359375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0205078125,
          "residual_disagree": 0.068359375
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.0205078125,
          "residual_disagree": 0.068359375
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.053059895833333336,
          "residual_disagree": 0.05322265625
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059244791666666664,
          "residual_disagree": 0.05029296875
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059244791666666664,
          "residual_disagree": 0.05029296875
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059244791666666664,
          "residual_disagree": 0.05029296875
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059244791666666664,
          "residual_disagree": 0.05029296875
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059244791666666664,
          "residual_disagree": 0.05029296875
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.059244791666666664,
          "residual_disagree": 0.05029296875
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.08870442708333333,
          "residual_disagree": 0.038899739583333336
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09440104166666667,
          "residual_disagree": 0.037272135416666664
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09440104166666667,
          "residual_disagree": 0.037272135416666664
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09440104166666667,
          "residual_disagree": 0.037272135416666664
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09440104166666667,
          "residual_disagree": 0.037272135416666664
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.09440104166666667,
          "residual_disagree": 0.037272135416666664
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.11735026041666667,
          "residual_disagree": 0.029296875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12386067708333333,
          "residual_disagree": 0.02685546875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12386067708333333,
          "residual_disagree": 0.02685546875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12386067708333333,
          "residual_disagree": 0.02685546875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12386067708333333,
          "residual_disagree": 0.02685546875
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.12386067708333333,
          "residual_disagree": 0.02685546875
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.14680989583333334,
          "residual_disagree": 0.02197265625
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.15185546875,
          "residual_disagree": 0.020670572916666668
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.15185546875,
          "residual_disagree": 0.020670572916666668
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.15185546875,
          "residual_disagree": 0.020670572916666668
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.15185546875,
          "residual_disagree": 0.020670572916666668
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.17366536458333334,
          "residual_disagree": 0.016764322916666668
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.17708333333333334,
          "residual_disagree": 0.016276041666666668
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.17708333333333334,
          "residual_disagree": 0.016276041666666668
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.17708333333333334,
          "residual_disagree": 0.016276041666666668
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.17708333333333334,
          "residual_disagree": 0.016276041666666668
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.197265625,
          "residual_disagree": 0.012858072916666666
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.201171875,
          "residual_disagree": 0.01220703125
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.201171875,
          "residual_disagree": 0.01220703125
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.201171875,
          "residual_disagree": 0.01220703125
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.201171875,
          "residual_disagree": 0.01220703125
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.22607421875,
          "residual_disagree": 0.009928385416666666
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.22607421875,
          "residual_disagree": 0.009928385416666666
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.22607421875,
          "residual_disagree": 0.009928385416666666
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.22607421875,
          "residual_disagree": 0.009928385416666666
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.24527994791666666,
          "residual_disagree": 0.008626302083333334
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.25,
          "residual_disagree": 0.008463541666666666
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.25,
          "residual_disagree": 0.008463541666666666
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.25,
          "residual_disagree": 0.008463541666666666
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.25,
          "residual_disagree": 0.008463541666666666
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.26953125,
          "residual_disagree": 0.00634765625
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.275390625,
          "residual_disagree": 0.006184895833333333
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.275390625,
          "residual_disagree": 0.006184895833333333
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.275390625,
          "residual_disagree": 0.006184895833333333
        },
        {
          "tau": 2.625,
          "escalated_frac": 0.291015625,
          "residual_disagree": 0.00537109375
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.2950846354166667,
          "residual_disagree": 0.005208333333333333
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.2950846354166667,
          "residual_disagree": 0.005208333333333333
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.2950846354166667,
          "residual_disagree": 0.005208333333333333
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.2950846354166667,
          "residual_disagree": 0.005208333333333333
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3147786458333333,
          "residual_disagree": 0.00390625
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3147786458333333,
          "residual_disagree": 0.00390625
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3147786458333333,
          "residual_disagree": 0.00390625
        },
        {
          "tau": 3.125,
          "escalated_frac": 0.3308919270833333,
          "residual_disagree": 0.0030924479166666665
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.3356119791666667,
          "residual_disagree": 0.0030924479166666665
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.3356119791666667,
          "residual_disagree": 0.0030924479166666665
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.3356119791666667,
          "residual_disagree": 0.0030924479166666665
        },
        {
          "tau": 3.375,
          "escalated_frac": 0.3505859375,
          "residual_disagree": 0.0022786458333333335
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.3541666666666667,
          "residual_disagree": 0.0021158854166666665
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.3541666666666667,
          "residual_disagree": 0.0021158854166666665
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.3541666666666667,
          "residual_disagree": 0.0021158854166666665
        },
        {
          "tau": 3.625,
          "escalated_frac": 0.3693033854166667,
          "residual_disagree": 0.0016276041666666667
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.3727213541666667,
          "residual_disagree": 0.0016276041666666667
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.3727213541666667,
          "residual_disagree": 0.0016276041666666667
        },
        {
          "tau": 3.875,
          "escalated_frac": 0.3841145833333333,
          "residual_disagree": 0.00146484375
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.3875325520833333,
          "residual_disagree": 0.00146484375
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.3875325520833333,
          "residual_disagree": 0.00146484375
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.4007161458333333,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.4007161458333333,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.4007161458333333,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.416015625,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.416015625,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.416015625,
          "residual_disagree": 0.0009765625
        },
        {
          "tau": 4.625,
          "escalated_frac": 0.4305013020833333,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.43310546875,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.43310546875,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.4449869791666667,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.4449869791666667,
          "residual_disagree": 0.0006510416666666666
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.4568684895833333,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.4568684895833333,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 5.375,
          "escalated_frac": 0.4664713541666667,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.46826171875,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.46826171875,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.48046875,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.48046875,
          "residual_disagree": 0.00048828125
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.4910481770833333,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.4910481770833333,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 6.125,
          "escalated_frac": 0.5009765625,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.5026041666666666,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 6.375,
          "escalated_frac": 0.51025390625,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 6.5,
          "escalated_frac": 0.5130208333333334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 6.625,
          "escalated_frac": 0.52099609375,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 6.75,
          "escalated_frac": 0.5234375,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 7.0,
          "escalated_frac": 0.5305989583333334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 7.0,
          "escalated_frac": 0.5305989583333334,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 7.25,
          "escalated_frac": 0.541015625,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 7.25,
          "escalated_frac": 0.541015625,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 7.5,
          "escalated_frac": 0.55029296875,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 7.625,
          "escalated_frac": 0.5568033854166666,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 7.75,
          "escalated_frac": 0.5589192708333334,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 7.875,
          "escalated_frac": 0.5673828125,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 8.0,
          "escalated_frac": 0.5690104166666666,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 8.25,
          "escalated_frac": 0.5763346354166666,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 8.375,
          "escalated_frac": 0.58203125,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 8.5,
          "escalated_frac": 0.58349609375,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 8.75,
          "escalated_frac": 0.5908203125,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 8.875,
          "escalated_frac": 0.595703125,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 9.0,
          "escalated_frac": 0.5983072916666666,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 9.25,
          "escalated_frac": 0.6064453125,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 9.375,
          "escalated_frac": 0.6111653645833334,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 9.5,
          "escalated_frac": 0.6139322916666666,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 9.75,
          "escalated_frac": 0.62060546875,
          "residual_disagree": 0.00016276041666666666
        },
        {
          "tau": 10.0,
          "escalated_frac": 0.6271158854166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.0,
          "escalated_frac": 0.6271158854166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.25,
          "escalated_frac": 0.6355794270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.5,
          "escalated_frac": 0.64306640625,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.5,
          "escalated_frac": 0.64306640625,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.75,
          "escalated_frac": 0.6510416666666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.0,
          "escalated_frac": 0.6564127604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.25,
          "escalated_frac": 0.66259765625,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.375,
          "escalated_frac": 0.6673177083333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.5,
          "escalated_frac": 0.6695963541666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.75,
          "escalated_frac": 0.6754557291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.0,
          "escalated_frac": 0.6814778645833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.25,
          "escalated_frac": 0.6871744791666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.25,
          "escalated_frac": 0.6871744791666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.5,
          "escalated_frac": 0.6956380208333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.75,
          "escalated_frac": 0.7021484375,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.875,
          "escalated_frac": 0.70703125,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.125,
          "escalated_frac": 0.7130533854166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.25,
          "escalated_frac": 0.7137044270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.5,
          "escalated_frac": 0.7210286458333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.75,
          "escalated_frac": 0.7275390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.0,
          "escalated_frac": 0.7327473958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.125,
          "escalated_frac": 0.73779296875,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.375,
          "escalated_frac": 0.7425130208333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.5,
          "escalated_frac": 0.744140625,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.75,
          "escalated_frac": 0.75146484375,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.0,
          "escalated_frac": 0.75830078125,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.125,
          "escalated_frac": 0.7618815104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.375,
          "escalated_frac": 0.7682291666666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.75,
          "escalated_frac": 0.7734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.875,
          "escalated_frac": 0.7766927083333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.0,
          "escalated_frac": 0.77978515625,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.25,
          "escalated_frac": 0.7859700520833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.625,
          "escalated_frac": 0.7931315104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.875,
          "escalated_frac": 0.7980143229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.0,
          "escalated_frac": 0.8004557291666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.375,
          "escalated_frac": 0.8084309895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.625,
          "escalated_frac": 0.8126627604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.875,
          "escalated_frac": 0.81787109375,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.125,
          "escalated_frac": 0.8232421875,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.375,
          "escalated_frac": 0.8284505208333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.625,
          "escalated_frac": 0.8330078125,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.75,
          "escalated_frac": 0.83544921875,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.0,
          "escalated_frac": 0.8421223958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.25,
          "escalated_frac": 0.8479817708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.375,
          "escalated_frac": 0.8523763020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.625,
          "escalated_frac": 0.8577473958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.0,
          "escalated_frac": 0.8631184895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.25,
          "escalated_frac": 0.8689778645833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.375,
          "escalated_frac": 0.8727213541666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.625,
          "escalated_frac": 0.8790690104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.0,
          "escalated_frac": 0.88427734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.125,
          "escalated_frac": 0.8885091145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.4839824120603,
          "escalated_frac": 0.8943684895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.75,
          "escalated_frac": 0.8987630208333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.0,
          "escalated_frac": 0.9015299479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.5,
          "escalated_frac": 0.9083658854166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.77763819095469,
          "escalated_frac": 0.91455078125,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.25,
          "escalated_frac": 0.91943359375,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.5,
          "escalated_frac": 0.9226888020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.875,
          "escalated_frac": 0.9288736979166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.25,
          "escalated_frac": 0.9329427083333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.5,
          "escalated_frac": 0.9366861979166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.0,
          "escalated_frac": 0.9441731770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.375,
          "escalated_frac": 0.9493815104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.75,
          "escalated_frac": 0.9534505208333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 26.25,
          "escalated_frac": 0.9581705729166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 26.875,
          "escalated_frac": 0.96435546875,
          "residual_disagree": 0.0
        },
        {
          "tau": 27.472989949748694,
          "escalated_frac": 0.9697265625,
          "residual_disagree": 0.0
        },
        {
          "tau": 28.25,
          "escalated_frac": 0.9742838541666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 29.25,
          "escalated_frac": 0.9793294270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 30.25,
          "escalated_frac": 0.9847005208333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 31.375,
          "escalated_frac": 0.9895833333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 32.875,
          "escalated_frac": 0.9947916666666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 38.375,
          "escalated_frac": 0.9998372395833334,
          "residual_disagree": 0.0
        }
      ],
      "per_domain": {
        "code": {
          "agreement_rate": 0.923828125,
          "auroc": 0.9159416165230119
        },
        "math": {
          "agreement_rate": 0.9462890625,
          "auroc": 0.9371986115020171
        },
        "chat": {
          "agreement_rate": 0.896484375,
          "auroc": 0.9169492744686973
        },
        "french": {
          "agreement_rate": 0.91796875,
          "auroc": 0.9131585612968591
        },
        "english": {
          "agreement_rate": 0.90234375,
          "auroc": 0.9038365800865801
        },
        "reasoning": {
          "agreement_rate": 0.9404296875,
          "auroc": 0.94566161074511
        }
      },
      "escalation_frac_for_99pct": 0.22607421875
    },
    "8": {
      "bits": 8,
      "n_positions": 6144,
      "agreement_rate": 0.9866536458333334,
      "mean_kl_ref_q": 0.0017304812790825963,
      "auroc_margin_predicts_disagreement": 0.980641098888719,
      "median_margin_agree": 6.5,
      "median_margin_disagree": 0.125,
      "escalation_curve": [
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.013346354166666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.013346354166666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.013346354166666666
        },
        {
          "tau": 0.0,
          "escalated_frac": 0.0,
          "residual_disagree": 0.013346354166666666
        },
        {
          "tau": 0.125,
          "escalated_frac": 0.015950520833333332,
          "residual_disagree": 0.008138020833333334
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.02099609375,
          "residual_disagree": 0.006510416666666667
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.02099609375,
          "residual_disagree": 0.006510416666666667
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.02099609375,
          "residual_disagree": 0.006510416666666667
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.02099609375,
          "residual_disagree": 0.006510416666666667
        },
        {
          "tau": 0.25,
          "escalated_frac": 0.02099609375,
          "residual_disagree": 0.006510416666666667
        },
        {
          "tau": 0.375,
          "escalated_frac": 0.047037760416666664,
          "residual_disagree": 0.00146484375
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.052408854166666664,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.052408854166666664,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.052408854166666664,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.052408854166666664,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.5,
          "escalated_frac": 0.052408854166666664,
          "residual_disagree": 0.0011393229166666667
        },
        {
          "tau": 0.625,
          "escalated_frac": 0.07845052083333333,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08463541666666667,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08463541666666667,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08463541666666667,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08463541666666667,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 0.75,
          "escalated_frac": 0.08463541666666667,
          "residual_disagree": 0.0003255208333333333
        },
        {
          "tau": 0.875,
          "escalated_frac": 0.10774739583333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.11246744791666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.11246744791666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.11246744791666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.0,
          "escalated_frac": 0.11246744791666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.125,
          "escalated_frac": 0.13509114583333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.140625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.140625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.140625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.140625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.25,
          "escalated_frac": 0.140625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.375,
          "escalated_frac": 0.16145833333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.16650390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.16650390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.16650390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.5,
          "escalated_frac": 0.16650390625,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.625,
          "escalated_frac": 0.1865234375,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.19156901041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.19156901041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.75,
          "escalated_frac": 0.19156901041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 1.875,
          "escalated_frac": 0.21012369791666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.21468098958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.21468098958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.21468098958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.0,
          "escalated_frac": 0.21468098958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.125,
          "escalated_frac": 0.23583984375,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.23974609375,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.23974609375,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.23974609375,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.25,
          "escalated_frac": 0.23974609375,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.375,
          "escalated_frac": 0.2579752604166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2618815104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2618815104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.5,
          "escalated_frac": 0.2618815104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.28125,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.28125,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.28125,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.75,
          "escalated_frac": 0.28125,
          "residual_disagree": 0.0
        },
        {
          "tau": 2.875,
          "escalated_frac": 0.2985026041666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3028971354166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.0,
          "escalated_frac": 0.3028971354166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.096105527638173,
          "escalated_frac": 0.3165690104166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.3196614583333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.3196614583333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.25,
          "escalated_frac": 0.3196614583333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.375,
          "escalated_frac": 0.3361002604166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.3385416666666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.3385416666666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.5,
          "escalated_frac": 0.3385416666666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.3557942708333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.3557942708333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 3.75,
          "escalated_frac": 0.3557942708333333,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.37158203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.37158203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.0,
          "escalated_frac": 0.37158203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.38427734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.25,
          "escalated_frac": 0.38427734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.39501953125,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.5,
          "escalated_frac": 0.39501953125,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.625,
          "escalated_frac": 0.4052734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.40771484375,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.75,
          "escalated_frac": 0.40771484375,
          "residual_disagree": 0.0
        },
        {
          "tau": 4.875,
          "escalated_frac": 0.4207356770833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.423828125,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.0,
          "escalated_frac": 0.423828125,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.125,
          "escalated_frac": 0.43603515625,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.4386393229166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.25,
          "escalated_frac": 0.4386393229166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.375,
          "escalated_frac": 0.44970703125,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.453125,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.5,
          "escalated_frac": 0.453125,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.4666341145833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.75,
          "escalated_frac": 0.4666341145833333,
          "residual_disagree": 0.0
        },
        {
          "tau": 5.875,
          "escalated_frac": 0.4762369791666667,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.4786783854166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.0,
          "escalated_frac": 0.4786783854166667,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.49169921875,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.25,
          "escalated_frac": 0.49169921875,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.5,
          "escalated_frac": 0.4990234375,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.5,
          "escalated_frac": 0.4990234375,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.75,
          "escalated_frac": 0.5104166666666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 6.75,
          "escalated_frac": 0.5104166666666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.0,
          "escalated_frac": 0.5193684895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.25,
          "escalated_frac": 0.5271809895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.25,
          "escalated_frac": 0.5271809895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.375,
          "escalated_frac": 0.5359700520833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.5,
          "escalated_frac": 0.5380859375,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.75,
          "escalated_frac": 0.5462239583333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 7.953517587939757,
          "escalated_frac": 0.552734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.0,
          "escalated_frac": 0.552734375,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.25,
          "escalated_frac": 0.5616861979166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.25,
          "escalated_frac": 0.5616861979166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.5,
          "escalated_frac": 0.56982421875,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.625,
          "escalated_frac": 0.576171875,
          "residual_disagree": 0.0
        },
        {
          "tau": 8.75,
          "escalated_frac": 0.5779622395833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.0,
          "escalated_frac": 0.5847981770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.25,
          "escalated_frac": 0.5914713541666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.5,
          "escalated_frac": 0.5978190104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.5,
          "escalated_frac": 0.5978190104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 9.75,
          "escalated_frac": 0.6046549479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.0,
          "escalated_frac": 0.61279296875,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.24120603015075,
          "escalated_frac": 0.6180013020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.25,
          "escalated_frac": 0.6180013020833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.5,
          "escalated_frac": 0.625,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.75,
          "escalated_frac": 0.6321614583333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 10.875,
          "escalated_frac": 0.6370442708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.0,
          "escalated_frac": 0.638671875,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.25,
          "escalated_frac": 0.6476236979166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.5,
          "escalated_frac": 0.6529947916666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.610552763819044,
          "escalated_frac": 0.658203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 11.75,
          "escalated_frac": 0.65966796875,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.0,
          "escalated_frac": 0.6681315104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.25,
          "escalated_frac": 0.6730143229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.375,
          "escalated_frac": 0.6774088541666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.5,
          "escalated_frac": 0.6796875,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.75,
          "escalated_frac": 0.6875,
          "residual_disagree": 0.0
        },
        {
          "tau": 12.875,
          "escalated_frac": 0.6917317708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.0,
          "escalated_frac": 0.6941731770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.25,
          "escalated_frac": 0.7014973958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.375,
          "escalated_frac": 0.7073567708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.5,
          "escalated_frac": 0.7097981770833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 13.75,
          "escalated_frac": 0.7169596354166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.0,
          "escalated_frac": 0.7228190104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.125,
          "escalated_frac": 0.72705078125,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.375,
          "escalated_frac": 0.7333984375,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.5,
          "escalated_frac": 0.7353515625,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.75,
          "escalated_frac": 0.7433268229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 14.941582914572905,
          "escalated_frac": 0.7486979166666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.125,
          "escalated_frac": 0.7522786458333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.375,
          "escalated_frac": 0.7584635416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.625,
          "escalated_frac": 0.7628580729166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 15.875,
          "escalated_frac": 0.7683919270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.0,
          "escalated_frac": 0.7706705729166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.25,
          "escalated_frac": 0.7775065104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.5,
          "escalated_frac": 0.7828776041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 16.75,
          "escalated_frac": 0.7884114583333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.0,
          "escalated_frac": 0.7937825520833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.125,
          "escalated_frac": 0.796875,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.25,
          "escalated_frac": 0.7998046875,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.625,
          "escalated_frac": 0.8084309895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 17.875,
          "escalated_frac": 0.8116861979166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.125,
          "escalated_frac": 0.8185221354166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.321608040201,
          "escalated_frac": 0.8240559895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.625,
          "escalated_frac": 0.8287760416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 18.75,
          "escalated_frac": 0.8304036458333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.125,
          "escalated_frac": 0.8390299479166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.25,
          "escalated_frac": 0.8416341145833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.5,
          "escalated_frac": 0.8483072916666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 19.75,
          "escalated_frac": 0.8527018229166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.0,
          "escalated_frac": 0.85791015625,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.25,
          "escalated_frac": 0.8636067708333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.5,
          "escalated_frac": 0.8673502604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 20.875,
          "escalated_frac": 0.8723958333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.25,
          "escalated_frac": 0.87890625,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.5,
          "escalated_frac": 0.8837890625,
          "residual_disagree": 0.0
        },
        {
          "tau": 21.75,
          "escalated_frac": 0.8880208333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.0,
          "escalated_frac": 0.8929036458333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.25,
          "escalated_frac": 0.8976236979166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.625,
          "escalated_frac": 0.904296875,
          "residual_disagree": 0.0
        },
        {
          "tau": 22.875,
          "escalated_frac": 0.908203125,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.125,
          "escalated_frac": 0.9134114583333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.5,
          "escalated_frac": 0.9178059895833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 23.75,
          "escalated_frac": 0.9236653645833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.0,
          "escalated_frac": 0.9280598958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.25,
          "escalated_frac": 0.9322916666666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 24.75,
          "escalated_frac": 0.939453125,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.0,
          "escalated_frac": 0.9436848958333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.5,
          "escalated_frac": 0.9493815104166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 25.771984924623098,
          "escalated_frac": 0.9547526041666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 26.375,
          "escalated_frac": 0.9596354166666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 27.125,
          "escalated_frac": 0.96435546875,
          "residual_disagree": 0.0
        },
        {
          "tau": 27.75,
          "escalated_frac": 0.9694010416666666,
          "residual_disagree": 0.0
        },
        {
          "tau": 28.375,
          "escalated_frac": 0.9739583333333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 29.315326633165796,
          "escalated_frac": 0.9798177083333334,
          "residual_disagree": 0.0
        },
        {
          "tau": 30.375,
          "escalated_frac": 0.9845377604166666,
          "residual_disagree": 0.0
        },
        {
          "tau": 31.375,
          "escalated_frac": 0.9892578125,
          "residual_disagree": 0.0
        },
        {
          "tau": 33.14133165829139,
          "escalated_frac": 0.9949544270833334,
          "residual_disagree": 0.0
        },
        {
          "tau": 39.0625,
          "escalated_frac": 0.9998372395833334,
          "residual_disagree": 0.0
        }
      ],
      "per_domain": {
        "code": {
          "agreement_rate": 0.98828125,
          "auroc": 0.9835309617918313
        },
        "math": {
          "agreement_rate": 0.9921875,
          "auroc": 0.9931102362204725
        },
        "chat": {
          "agreement_rate": 0.9765625,
          "auroc": 0.967625
        },
        "french": {
          "agreement_rate": 0.98828125,
          "auroc": 0.9731966403162056
        },
        "english": {
          "agreement_rate": 0.982421875,
          "auroc": 0.9842611000662691
        },
        "reasoning": {
          "agreement_rate": 0.9921875,
          "auroc": 0.9833907480314961
        }
      },
      "escalation_frac_for_99pct": 0.015950520833333332
    }
  }
}