The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
baseline_metrics: struct<dev: struct<cross_coherence: struct<auprc: double, auroc: double>, dealt_ewma: struct<auprc: (... 1397 chars omitted)
child 0, dev: struct<cross_coherence: struct<auprc: double, auroc: double>, dealt_ewma: struct<auprc: double, auro (... 388 chars omitted)
child 0, cross_coherence: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 1, dealt_ewma: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 2, energy_imbalance: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 3, hit_recency: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 4, phase_alignment: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 5, received_energy: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 6, received_ewma: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 7, recent_dealt: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 8, recent_received: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 1, ho
...
oc: double>, holdout: s (... 95 chars omitted)
child 0, best_dev_auprc: double
child 1, metrics: struct<dev: struct<auprc: double, auroc: double>, holdout: struct<auprc: double, auroc: double>, tra (... 41 chars omitted)
child 0, dev: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 1, holdout: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 2, train: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 2, seed: int64
status: string
wandb_runs: list<item: string>
child 0, item: string
retrospective_holdout_seeds: list<item: int64>
child 0, item: int64
taus: list<item: double>
child 0, item: double
note: string
horizon_frames: int64
periods: list<item: double>
child 0, item: double
source_manifest_sha256: string
samples: struct<dev: int64, holdout: int64, train: int64>
child 0, dev: int64
child 1, holdout: int64
child 2, train: int64
sample_frames: list<item: int64>
child 0, item: int64
source_commit: string
source_monitor_events_sha256: string
dev_seeds: list<item: int64>
child 0, item: int64
train_seeds: list<item: int64>
child 0, item: int64
source_matches_sha256: string
positive_rate: struct<dev: double, holdout: double, train: double>
child 0, dev: double
child 1, holdout: double
child 2, train: double
to
{'dev_seeds': List(Value('int64')), 'horizon_frames': Value('int64'), 'note': Value('string'), 'periods': List(Value('float64')), 'positive_rate': {'dev': Value('float64'), 'holdout': Value('float64'), 'train': Value('float64')}, 'retrospective_holdout_seeds': List(Value('int64')), 'sample_frames': List(Value('int64')), 'samples': {'dev': Value('int64'), 'holdout': Value('int64'), 'train': Value('int64')}, 'source_commit': Value('string'), 'source_manifest_sha256': Value('string'), 'source_matches_sha256': Value('string'), 'source_monitor_events_sha256': Value('string'), 'taus': List(Value('float64')), 'train_seeds': List(Value('int64'))}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
return get_rows(
dataset=dataset,
...<4 lines>...
column_names=column_names,
)
File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
return func(*args, **kwargs)
File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
yield from ds.decode(False) if ds.features else ds
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2840, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, in _iter_arrow
for key, pa_table in self.ex_iterable._iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
baseline_metrics: struct<dev: struct<cross_coherence: struct<auprc: double, auroc: double>, dealt_ewma: struct<auprc: (... 1397 chars omitted)
child 0, dev: struct<cross_coherence: struct<auprc: double, auroc: double>, dealt_ewma: struct<auprc: double, auro (... 388 chars omitted)
child 0, cross_coherence: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 1, dealt_ewma: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 2, energy_imbalance: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 3, hit_recency: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 4, phase_alignment: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 5, received_energy: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 6, received_ewma: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 7, recent_dealt: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 8, recent_received: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 1, ho
...
oc: double>, holdout: s (... 95 chars omitted)
child 0, best_dev_auprc: double
child 1, metrics: struct<dev: struct<auprc: double, auroc: double>, holdout: struct<auprc: double, auroc: double>, tra (... 41 chars omitted)
child 0, dev: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 1, holdout: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 2, train: struct<auprc: double, auroc: double>
child 0, auprc: double
child 1, auroc: double
child 2, seed: int64
status: string
wandb_runs: list<item: string>
child 0, item: string
retrospective_holdout_seeds: list<item: int64>
child 0, item: int64
taus: list<item: double>
child 0, item: double
note: string
horizon_frames: int64
periods: list<item: double>
child 0, item: double
source_manifest_sha256: string
samples: struct<dev: int64, holdout: int64, train: int64>
child 0, dev: int64
child 1, holdout: int64
child 2, train: int64
sample_frames: list<item: int64>
child 0, item: int64
source_commit: string
source_monitor_events_sha256: string
dev_seeds: list<item: int64>
child 0, item: int64
train_seeds: list<item: int64>
child 0, item: int64
source_matches_sha256: string
positive_rate: struct<dev: double, holdout: double, train: double>
child 0, dev: double
child 1, holdout: double
child 2, train: double
to
{'dev_seeds': List(Value('int64')), 'horizon_frames': Value('int64'), 'note': Value('string'), 'periods': List(Value('float64')), 'positive_rate': {'dev': Value('float64'), 'holdout': Value('float64'), 'train': Value('float64')}, 'retrospective_holdout_seeds': List(Value('int64')), 'sample_frames': List(Value('int64')), 'samples': {'dev': Value('int64'), 'holdout': Value('int64'), 'train': Value('int64')}, 'source_commit': Value('string'), 'source_manifest_sha256': Value('string'), 'source_matches_sha256': Value('string'), 'source_monitor_events_sha256': Value('string'), 'taus': List(Value('float64')), 'train_seeds': List(Value('int64'))}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
SSH Fighter Combat-Kick Oscillator v1
Passive offline detector evidence over the validated exact-engine Gym Leg-2 matrix. Authoritative HP-loss events are literal impulses into eight fixed complex oscillator modes. The modes are never trained; three tiny linear readouts predict whether the observed side will receive damage in the next 30 frames.
The grouped retrospective holdout result is AUROC 0.75796 and AUPRC 0.47426 (positive prevalence 0.26093). The strongest cheap causal baseline, received damage EWMA, is AUROC 0.69470 / AUPRC 0.36525.
Important limits:
- This uses exact-engine HP changes, not server hit/combo telemetry.
- HP-loss impulses are not kick-button inputs.
- Source snapshot
f478c1320f86f428c882e391c11e97ca4cf7b47bis byte-identical to canonical main3caedf3435c12996cf4d34fb5ac76c7cd7b75076across engine, moves, types, roster, bot wire, and replay simulation, but the episodes retain their historical pre-intervention epoch label. - The holdout seeds were previously visible in aggregate and are not sealed.
- No live transport, policy mutation, or balance intervention is represented.
combat_kick_dataset.npz contains the causal samples and monitor values;
checkpoint_seed*.pt are the three readout-only checkpoints; dataset.json
and results.json contain provenance and full metrics.
- Downloads last month
- 51