tier 1 · fault_grounded · global view · 2 classes · all cells measured
Headline metric balanced_accuracy (↑ higher is better); full metric set: balanced_accuracy · macro_f1 · accuracy. Metrics are a frozen contract (src/evemtbench/evaluation/metrics.py); label derivation: src/evemtbench/tasks/labels.
event flt_1phg_shc_w_arc · grid double_line (adapt_grid/test split) · sample 29, window #667, cubicle 3 (max-RMS) · 50 ms / 480 samples @ 9.6 kHz · top: 3-phase current, bottom: 3-phase voltage · figure provenance in assets/manifest.json
| task | grid | baseline | headline | test | benchmark | params | fit/seed | trace |
|---|---|---|---|---|---|---|---|---|
| fault_grounded_global | cigre_mv | majority | balanced_accuracy ↑ | 0.500 ± 0.000 (n=5) | 0.500 ± 0.000 (n=5) | — | 0s | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | cigre_mv | random_forest | balanced_accuracy ↑ | 0.935 ± 0.002 (n=5) | 0.989 ± 0.001 (n=5) | — | 1.1m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | cigre_mv | mlp | balanced_accuracy ↑ | 0.928 ± 0.013 (n=5) | 0.975 ± 0.008 (n=5) | 42.9M | 7.0m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | cigre_mv | gru | balanced_accuracy ↑ | 0.944 ± 0.012 (n=5) | 0.986 ± 0.005 (n=5) | 117k | 5.9m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | cigre_mv | cnn | balanced_accuracy ↑ | 0.946 ± 0.007 (n=5) | 0.988 ± 0.005 (n=5) | 124k | 6.8m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | cigre_mv | resnet | balanced_accuracy ↑ | 0.953 ± 0.008 (n=5) | 0.994 ± 0.002 (n=5) | 602k | 6.7m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | double_line | majority | balanced_accuracy ↑ | 0.500 ± 0.000 (n=5) | 0.500 ± 0.000 (n=5) | — | 0s | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | double_line | random_forest | balanced_accuracy ↑ | 0.955 ± 0.000 (n=5) | 0.653 ± 0.007 (n=5) | — | 21s | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | double_line | mlp | balanced_accuracy ↑ | 0.969 ± 0.000 (n=5) | 0.998 ± 0.001 (n=5) | 12.0M | 3.3m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | double_line | gru | balanced_accuracy ↑ | 0.970 ± 0.000 (n=5) | 0.999 ± 0.001 (n=5) | 69k | 2.5m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | double_line | cnn | balanced_accuracy ↑ | 0.971 ± 0.000 (n=5) | 1.000 ± 0.001 (n=5) | 68k | 3.2m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | double_line | resnet | balanced_accuracy ↑ | 0.970 ± 0.000 (n=5) | 0.999 ± 0.001 (n=5) | 530k | 6.1m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | testgrid_110kv | majority | balanced_accuracy ↑ | 0.500 ± 0.000 (n=5) | 0.500 ± 0.000 (n=5) | — | 0s | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | testgrid_110kv | random_forest | balanced_accuracy ↑ | 0.960 ± 0.001 (n=5) | 0.680 ± 0.008 (n=5) | — | 36s | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | testgrid_110kv | mlp | balanced_accuracy ↑ | 0.969 ± 0.000 (n=5) | 0.995 ± 0.000 (n=5) | 26.7M | 6.3m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | testgrid_110kv | gru | balanced_accuracy ↑ | 0.970 ± 0.000 (n=5) | 0.996 ± 0.002 (n=5) | 92k | 4.4m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | testgrid_110kv | cnn | balanced_accuracy ↑ | 0.969 ± 0.001 (n=5) | 0.999 ± 0.001 (n=5) | 94k | 5.0m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | testgrid_110kv | resnet | balanced_accuracy ↑ | 0.969 ± 0.001 (n=5) | 0.998 ± 0.002 (n=5) | 564k | 4.4m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | ieee39 | majority | balanced_accuracy ↑ | 0.500 ± 0.000 (n=5) | 0.500 ± 0.000 (n=5) | — | 0s | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | ieee39 | random_forest | balanced_accuracy ↑ | 0.955 ± 0.001 (n=5) | 0.870 ± 0.009 (n=5) | — | 2.8m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | ieee39 | mlp | balanced_accuracy ↑ | 0.950 ± 0.003 (n=5) | 0.926 ± 0.006 (n=5) | 103.4M | 33.0m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | ieee39 | gru | balanced_accuracy ↑ | 0.936 ± 0.021 (n=5) | 0.902 ± 0.057 (n=5) | 211k | 22.9m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | ieee39 | cnn | balanced_accuracy ↑ | 0.967 ± 0.001 (n=5) | 0.961 ± 0.027 (n=5) | 234k | 30.0m | trace: output · eval-code · log · train-code · config · W&B |
| fault_grounded_global | ieee39 | resnet | balanced_accuracy ↑ | 0.967 ± 0.003 (n=5) | 0.967 ± 0.006 (n=5) | 744k | 27.2m | trace: output · eval-code · log · train-code · config · W&B |
headline & reported metrics for this view — the metric set is a frozen contract; each links to the exact function that computes it
| metric | definition | entry |
|---|---|---|
balanced_accuracy | mean of per-class recall — the majority class cannot buy a good score | evaluate() |
macro_f1 | unweighted mean of per-class F1 (zero_division=0) | evaluate() |
accuracy | fraction of exact-match predictions | evaluate() |
PYTHONPATH=src python -m evemtbench.baselines.runner --task fault_grounded_global --protocol held_out \
--grid <grid> --baseline <baseline> --seeds 0 1 2 3 4