Risk-Calibrated World-Model Gates / results.json

Failed on benchmark

Raw ⬇ ZIP
  1{
  2  "checks": {
  3    "max_exact_abs_error": 0.0027159743000001457,
  4    "max_probe_abs_error": 8.191216885759106e-05,
  5    "all_sizing_rows_meet_delta": true,
  6    "all_probe_rows_improve_over_baseline": true
  7  },
  8  "exact_miss_sweep": [
  9    {
 10      "r": 0.01,
 11      "N": 5,
 12      "observed_miss": 0.95172,
 13      "predicted_miss": 0.9509900498999999,
 14      "abs_error": 0.0007299501000000985
 15    },
 16    {
 17      "r": 0.01,
 18      "N": 20,
 19      "observed_miss": 0.81884,
 20      "predicted_miss": 0.8179069375972308,
 21      "abs_error": 0.0009330624027692469
 22    },
 23    {
 24      "r": 0.01,
 25      "N": 50,
 26      "observed_miss": 0.60504,
 27      "predicted_miss": 0.6050060671375364,
 28      "abs_error": 3.393286246367033e-05
 29    },
 30    {
 31      "r": 0.03,
 32      "N": 5,
 33      "observed_miss": 0.86145,
 34      "predicted_miss": 0.8587340256999999,
 35      "abs_error": 0.0027159743000001457
 36    },
 37    {
 38      "r": 0.03,
 39      "N": 20,
 40      "observed_miss": 0.54361,
 41      "predicted_miss": 0.543794342926747,
 42      "abs_error": 0.00018434292674696096
 43    },
 44    {
 45      "r": 0.03,
 46      "N": 50,
 47      "observed_miss": 0.21829,
 48      "predicted_miss": 0.2180653753474073,
 49      "abs_error": 0.0002246246525927087
 50    },
 51    {
 52      "r": 0.1,
 53      "N": 5,
 54      "observed_miss": 0.59122,
 55      "predicted_miss": 0.5904900000000001,
 56      "abs_error": 0.0007299999999998974
 57    },
 58    {
 59      "r": 0.1,
 60      "N": 20,
 61      "observed_miss": 0.12256,
 62      "predicted_miss": 0.12157665459056935,
 63      "abs_error": 0.0009833454094306526
 64    },
 65    {
 66      "r": 0.1,
 67      "N": 50,
 68      "observed_miss": 0.00551,
 69      "predicted_miss": 0.00515377520732012,
 70      "abs_error": 0.00035622479267988054
 71    },
 72    {
 73      "r": 0.25,
 74      "N": 5,
 75      "observed_miss": 0.23822,
 76      "predicted_miss": 0.2373046875,
 77      "abs_error": 0.0009153124999999873
 78    },
 79    {
 80      "r": 0.25,
 81      "N": 20,
 82      "observed_miss": 0.00325,
 83      "predicted_miss": 0.0031712119389339932,
 84      "abs_error": 7.878806106600663e-05
 85    },
 86    {
 87      "r": 0.25,
 88      "N": 50,
 89      "observed_miss": 0.0,
 90      "predicted_miss": 5.663216564269376e-07,
 91      "abs_error": 5.663216564269376e-07
 92    }
 93  ],
 94  "required_count_sweep": [
 95    {
 96      "r": 0.005,
 97      "N_required": 598,
 98      "exact_miss": 0.04991169123005841,
 99      "target_delta": 0.05,
100      "small_r_approx_N": 599.1464547107981,
101      "N_over_approx": 0.998086520079049
102    },
103    {
104      "r": 0.01,
105      "N_required": 299,
106      "exact_miss": 0.04953625663766235,
107      "target_delta": 0.05,
108      "small_r_approx_N": 299.57322735539907,
109      "N_over_approx": 0.998086520079049
110    },
111    {
112      "r": 0.02,
113      "N_required": 149,
114      "exact_miss": 0.04928165433008947,
115      "target_delta": 0.05,
116      "small_r_approx_N": 149.78661367769953,
117      "N_over_approx": 0.9947484380720956
118    },
119    {
120      "r": 0.05,
121      "N_required": 59,
122      "exact_miss": 0.048494525249423104,
123      "target_delta": 0.05,
124      "small_r_approx_N": 59.914645471079815,
125      "N_over_approx": 0.9847341920512356
126    },
127    {
128      "r": 0.1,
129      "N_required": 29,
130      "exact_miss": 0.047101286972462485,
131      "target_delta": 0.05,
132      "small_r_approx_N": 29.957322735539908,
133      "N_over_approx": 0.9680437820164688
134    }
135  ],
136  "matched_budget_probe_sweep": [
137    {
138      "r": 0.005,
139      "probe_rate_q": 0.15,
140      "baseline_miss_observed": 0.77877,
141      "baseline_miss_predicted": 0.7783125570686419,
142      "idea_miss_observed": 0.03343,
143      "idea_miss_predicted": 0.03334808783114241,
144      "detection_gain": -0.74534,
145      "probe_scaling_factor_q_over_r": 30.0
146    },
147    {
148      "r": 0.01,
149      "probe_rate_q": 0.3,
150      "baseline_miss_observed": 0.60664,
151      "baseline_miss_predicted": 0.6050060671375364,
152      "idea_miss_observed": 0.00065,
153      "idea_miss_predicted": 0.0005902236917384061,
154      "detection_gain": -0.6059899999999999,
155      "probe_scaling_factor_q_over_r": 30.0
156    },
157    {
158      "r": 0.02,
159      "probe_rate_q": 0.6,
160      "baseline_miss_observed": 0.36528,
161      "baseline_miss_predicted": 0.36416968008711675,
162      "idea_miss_observed": 0.0,
163      "idea_miss_predicted": 5.997663519304674e-09,
164      "detection_gain": -0.36528000000000005,
165      "probe_scaling_factor_q_over_r": 30.0
166    },
167    {
168      "r": 0.05,
169      "probe_rate_q": 0.6,
170      "baseline_miss_observed": 0.07742,
171      "baseline_miss_predicted": 0.07694497527671315,
172      "idea_miss_observed": 0.0,
173      "idea_miss_predicted": 2.3599781672672784e-09,
174      "detection_gain": -0.07742000000000004,
175      "probe_scaling_factor_q_over_r": 11.999999999999998
176    }
177  ]
178}