File size: 5,779 Bytes
5adf069 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 | {
"purpose": "Mechanics-only smoke validation, not improved released weights or generalization",
"adopted": false,
"training": {
"steps": 6,
"seed": 42,
"learningRate": 2e-05,
"trainRecords": 28,
"trainDomains": [
"garden",
"library",
"museum",
"workshop"
],
"changedAdapterTensors": 240,
"backboneAndReadoutFrozen": true,
"seconds": 53.129103966
},
"developmentBefore": {
"questions": 12,
"answered": 12,
"errors": 0,
"correct": 12,
"accuracyAllQuestions": 1.0,
"brierAnswered": 0.0010257946398990957,
"nllAnswered": 0.024988250780074998,
"ece10Answered": 0.024658157503599232,
"confidentWrongAt09": 0,
"coverageByFixedThreshold": [
{
"threshold": 0.5,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.7,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.8,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.9,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.95,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.99,
"accepted": 0,
"coverage": 0.0,
"errorRate": null
}
],
"coverageInterpretation": "Observed on this dataset; not a future error guarantee. Thresholds not tuned here."
},
"developmentAfter": {
"questions": 12,
"answered": 12,
"errors": 0,
"correct": 12,
"accuracyAllQuestions": 1.0,
"brierAnswered": 3.058330986685694e-05,
"nllAnswered": 0.004121613352848992,
"ece10Answered": 0.0041117529524682395,
"confidentWrongAt09": 0,
"coverageByFixedThreshold": [
{
"threshold": 0.5,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.7,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.8,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.9,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.95,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.99,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
}
],
"coverageInterpretation": "Observed on this dataset; not a future error guarantee. Thresholds not tuned here."
},
"reload": {
"complete": true,
"adopted": false,
"priorStrictGateFailed": true,
"all240AdapterTensorsBitExact": true,
"freshRepeat": {
"maxProbabilityDelta": 0.0,
"sameAllTopOptions": true,
"questions": 12
},
"crossProcessReload": {
"maxProbabilityDelta": 0.0,
"sameAllTopOptions": true,
"questions": 12
},
"originalStrictTolerancePassed": true,
"calibrationAndTestPassed": true,
"seconds": 196.0350929440001
},
"initialStrictReloadFailure": {
"maxProbabilityDelta": 0.00048164095202640755,
"tolerance": 1e-05,
"failed": true,
"interpretation": "Observed after compiled/HTTP work in the original process. Cause was not isolated. Preserved rather than erased by the successful fresh-process control."
},
"calibration": {
"records": 4,
"factor": 0.1353352832366127,
"factorBounds": [
0.1353352832366127,
7.38905609893065
],
"beforeNll": 0.003907485844979199,
"afterNll": 0.0,
"interpretation": "In-sample calibration fit; independent test needed before adoption",
"lowerBoundSelected": true
},
"toyTest": {
"questions": 12,
"answered": 12,
"errors": 0,
"correct": 12,
"accuracyAllQuestions": 1.0,
"brierAnswered": 3.9934072114968966e-38,
"nllAnswered": 0.0,
"ece10Answered": 0.0,
"confidentWrongAt09": 0,
"coverageByFixedThreshold": [
{
"threshold": 0.5,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.7,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.8,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.9,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.95,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
},
{
"threshold": 0.99,
"accepted": 12,
"coverage": 1.0,
"errorRate": 0.0
}
],
"coverageInterpretation": "Observed on this dataset; not a future error guarantee. Thresholds not tuned here."
},
"inputHashes": {
"train.jsonl": "cef745ddee9bf2e1d8f3639660441505bfb5aacf2c5cf9de5261a17ac1ba0bb5",
"development.jsonl": "c7da14026dbfd87b5dac9040c7b6ed33d3007b620aab018dc9c25ca5ce6d9059",
"calibration.jsonl": "b951c82c4263d6b0ec058b704f1d3d616ae85f483dc3d6dfdad49bed61c7481c",
"test.jsonl": "00baffcf3d9114d585294dd0955724ad9aff04475d3e9105584c9f1bbfbfbc8a"
},
"limitations": [
"Authored toy templates recur across partitions; this is not an independent cross-domain generalization test",
"Development was already 12/12 correct before training",
"Calibration chose its lower search bound on a tiny all-correct sample; this is not evidence of deployment calibration",
"No trained overlay or fitted temperature is adopted or included in the release"
]
}
|