| { | |
| "experiment": "P-MATH-INK-06-BOUNDARY-DEVICE-STRESS-001", | |
| "generated_at": "2026-07-23T19:53:29.966175+00:00", | |
| "device": "cuda", | |
| "cuda_device": "NVIDIA GeForce GTX 1650", | |
| "test_cache_key": "26bee2f320c6f0a7eca3c6ca1c9c71dd0cbe89cfaee8d8c415af74ae84c0d879", | |
| "stress_contract": { | |
| "model_variant": "base_plus_shared_adapter_plus_auxiliary_head", | |
| "modes": [ | |
| "clean", | |
| "coordinate_jitter", | |
| "timestamp_missing", | |
| "sparse_sampling", | |
| "affine_device" | |
| ], | |
| "maximum_exact_regression_pp": 3.0, | |
| "maximum_family_regression_pp": 3.0, | |
| "minimum_single_symbol_recall": 0.9, | |
| "minimum_cross_boundary_recall": 0.9 | |
| }, | |
| "seeds": [ | |
| { | |
| "seed": 17, | |
| "threshold": 0.725, | |
| "modes": { | |
| "clean": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8294553146483342, | |
| "family_top1": 0.9122157588577472 | |
| }, | |
| "boundary": { | |
| "threshold": 0.725, | |
| "accuracy": 0.9383333333333334, | |
| "f1": 0.9389438943894389, | |
| "roc_auc": 0.9841458333333333, | |
| "single_symbol_recall": 0.9283333333333333, | |
| "cross_boundary_recall": 0.9483333333333334 | |
| } | |
| }, | |
| "coordinate_jitter": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8313061872025383, | |
| "family_top1": 0.9127445795875199 | |
| }, | |
| "boundary": { | |
| "threshold": 0.725, | |
| "accuracy": 0.9383333333333334, | |
| "f1": 0.9389942291838417, | |
| "roc_auc": 0.9841597222222223, | |
| "single_symbol_recall": 0.9275, | |
| "cross_boundary_recall": 0.9491666666666667 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": 0.18508725542041304, | |
| "family_top1_pp": 0.052882072977267214 | |
| }, | |
| "gate_passed": true | |
| }, | |
| "timestamp_missing": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.825753569539926, | |
| "family_top1": 0.9106292966684294 | |
| }, | |
| "boundary": { | |
| "threshold": 0.725, | |
| "accuracy": 0.9366666666666666, | |
| "f1": 0.9372937293729373, | |
| "roc_auc": 0.9841229166666666, | |
| "single_symbol_recall": 0.9266666666666666, | |
| "cross_boundary_recall": 0.9466666666666667 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": -0.3701745108408261, | |
| "family_top1_pp": -0.15864621893177944 | |
| }, | |
| "gate_passed": true | |
| }, | |
| "sparse_sampling": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8220518244315177, | |
| "family_top1": 0.9085140137493389 | |
| }, | |
| "boundary": { | |
| "threshold": 0.725, | |
| "accuracy": 0.94125, | |
| "f1": 0.9420468557336621, | |
| "roc_auc": 0.9847597222222222, | |
| "single_symbol_recall": 0.9275, | |
| "cross_boundary_recall": 0.955 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": -0.7403490216816522, | |
| "family_top1_pp": -0.3701745108408261 | |
| }, | |
| "gate_passed": true | |
| }, | |
| "affine_device": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.819672131147541, | |
| "family_top1": 0.9013749338974087 | |
| }, | |
| "boundary": { | |
| "threshold": 0.725, | |
| "accuracy": 0.9308333333333333, | |
| "f1": 0.9310058187863675, | |
| "roc_auc": 0.979020138888889, | |
| "single_symbol_recall": 0.9283333333333333, | |
| "cross_boundary_recall": 0.9333333333333333 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": -0.9783183500793213, | |
| "family_top1_pp": -1.0840824960338447 | |
| }, | |
| "gate_passed": true | |
| } | |
| }, | |
| "all_stress_gates_passed": true | |
| }, | |
| { | |
| "seed": 31, | |
| "threshold": 0.7, | |
| "modes": { | |
| "clean": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8323638286620836, | |
| "family_top1": 0.9151242728714966 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.9604166666666667, | |
| "f1": 0.9610176446450553, | |
| "roc_auc": 0.9922055555555556, | |
| "single_symbol_recall": 0.945, | |
| "cross_boundary_recall": 0.9758333333333333 | |
| } | |
| }, | |
| "coordinate_jitter": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8323638286620836, | |
| "family_top1": 0.914595452141724 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.96, | |
| "f1": 0.9605911330049262, | |
| "roc_auc": 0.9922305555555556, | |
| "single_symbol_recall": 0.945, | |
| "cross_boundary_recall": 0.975 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": 0.0, | |
| "family_top1_pp": -0.05288207297725611 | |
| }, | |
| "gate_passed": true | |
| }, | |
| "timestamp_missing": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8291909042834479, | |
| "family_top1": 0.9103648863035431 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.96, | |
| "f1": 0.9605263157894737, | |
| "roc_auc": 0.9915847222222223, | |
| "single_symbol_recall": 0.9466666666666667, | |
| "cross_boundary_recall": 0.9733333333333334 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": -0.31729243786357, | |
| "family_top1_pp": -0.4759386567953494 | |
| }, | |
| "gate_passed": true | |
| }, | |
| "sparse_sampling": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8262823902696985, | |
| "family_top1": 0.9116869381279746 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.9579166666666666, | |
| "f1": 0.9586573884568154, | |
| "roc_auc": 0.9918708333333333, | |
| "single_symbol_recall": 0.94, | |
| "cross_boundary_recall": 0.9758333333333333 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": -0.6081438392385063, | |
| "family_top1_pp": -0.3437334743521925 | |
| }, | |
| "gate_passed": true | |
| }, | |
| "affine_device": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8276044420941301, | |
| "family_top1": 0.9069275515600211 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.9541666666666667, | |
| "f1": 0.9547325102880658, | |
| "roc_auc": 0.9886583333333334, | |
| "single_symbol_recall": 0.9416666666666667, | |
| "cross_boundary_recall": 0.9666666666666667 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": -0.4759386567953494, | |
| "family_top1_pp": -0.8196721311475419 | |
| }, | |
| "gate_passed": true | |
| } | |
| }, | |
| "all_stress_gates_passed": true | |
| }, | |
| { | |
| "seed": 47, | |
| "threshold": 0.7, | |
| "modes": { | |
| "clean": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8291909042834479, | |
| "family_top1": 0.9138022210470651 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.9408333333333333, | |
| "f1": 0.9423701298701299, | |
| "roc_auc": 0.9841138888888888, | |
| "single_symbol_recall": 0.9141666666666667, | |
| "cross_boundary_recall": 0.9675 | |
| } | |
| }, | |
| "coordinate_jitter": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8302485457429931, | |
| "family_top1": 0.9130089899524061 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.9404166666666667, | |
| "f1": 0.9419878296146045, | |
| "roc_auc": 0.9840753472222222, | |
| "single_symbol_recall": 0.9133333333333333, | |
| "cross_boundary_recall": 0.9675 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": 0.10576414595452333, | |
| "family_top1_pp": -0.07932310946590082 | |
| }, | |
| "gate_passed": true | |
| }, | |
| "timestamp_missing": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8270756213643575, | |
| "family_top1": 0.9151242728714966 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.9454166666666667, | |
| "f1": 0.9465087790935075, | |
| "roc_auc": 0.9864027777777777, | |
| "single_symbol_recall": 0.925, | |
| "cross_boundary_recall": 0.9658333333333333 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": -0.21152829190903555, | |
| "family_top1_pp": 0.13220518244314583 | |
| }, | |
| "gate_passed": true | |
| }, | |
| "sparse_sampling": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8254891591750396, | |
| "family_top1": 0.9122157588577472 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.9416666666666667, | |
| "f1": 0.9429967426710097, | |
| "roc_auc": 0.9858006944444445, | |
| "single_symbol_recall": 0.9183333333333333, | |
| "cross_boundary_recall": 0.965 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": -0.3701745108408261, | |
| "family_top1_pp": -0.15864621893179054 | |
| }, | |
| "gate_passed": true | |
| }, | |
| "affine_device": { | |
| "authentic": { | |
| "samples": 3782, | |
| "exact_top1": 0.8209941829719725, | |
| "family_top1": 0.9037546271813856 | |
| }, | |
| "boundary": { | |
| "threshold": 0.7, | |
| "accuracy": 0.9204166666666667, | |
| "f1": 0.9209109730848861, | |
| "roc_auc": 0.9774506944444444, | |
| "single_symbol_recall": 0.9141666666666667, | |
| "cross_boundary_recall": 0.9266666666666666 | |
| }, | |
| "deltas": { | |
| "exact_top1_pp": -0.8196721311475419, | |
| "family_top1_pp": -1.004759386567955 | |
| }, | |
| "gate_passed": true | |
| } | |
| }, | |
| "all_stress_gates_passed": true | |
| } | |
| ], | |
| "decision": { | |
| "all_seed_stress_gates_passed": true, | |
| "failures": [], | |
| "product_validation": false | |
| }, | |
| "track": "P_with_obligations", | |
| "product_validation": false | |
| } | |