Dataset Viewer
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
base_pos_rate: double
config: struct<do_bootstrap: bool, features_config_hash: string, n_boot: int64, probe_a_config_hash: string, (... 95 chars omitted)
child 0, do_bootstrap: bool
child 1, features_config_hash: string
child 2, n_boot: int64
child 3, probe_a_config_hash: string
child 4, probe_a_name: string
child 5, probe_b_config_hash: string
child 6, probe_b_name: string
child 7, scoring_mode: string
config_hash: string
n_test: int64
stage: string
systems: list<item: string>
child 0, item: string
timestamp_utc: string
split: string
n_samples: int64
n_available: int64
elapsed_s: double
judge_spec: string
scorers: list<item: string>
child 0, item: string
headline: struct<factuality: double>
child 0, factuality: double
to
{'elapsed_s': Value('float64'), 'headline': {'factuality': Value('float64')}, 'judge_spec': Value('string'), 'n_available': Value('int64'), 'n_samples': Value('int64'), 'scorers': List(Value('string')), 'split': Value('string')}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 149, 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 129, 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 544, 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 3052, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2559, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2584, 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 558, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 440, 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 361, 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 2381, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2309, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
base_pos_rate: double
config: struct<do_bootstrap: bool, features_config_hash: string, n_boot: int64, probe_a_config_hash: string, (... 95 chars omitted)
child 0, do_bootstrap: bool
child 1, features_config_hash: string
child 2, n_boot: int64
child 3, probe_a_config_hash: string
child 4, probe_a_name: string
child 5, probe_b_config_hash: string
child 6, probe_b_name: string
child 7, scoring_mode: string
config_hash: string
n_test: int64
stage: string
systems: list<item: string>
child 0, item: string
timestamp_utc: string
split: string
n_samples: int64
n_available: int64
elapsed_s: double
judge_spec: string
scorers: list<item: string>
child 0, item: string
headline: struct<factuality: double>
child 0, factuality: double
to
{'elapsed_s': Value('float64'), 'headline': {'factuality': Value('float64')}, 'judge_spec': Value('string'), 'n_available': Value('int64'), 'n_samples': Value('int64'), 'scorers': List(Value('string')), 'split': Value('string')}
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.
Tiny Aya Hallucination Guard — Results
Risk–coverage evaluation and SAE alignment for the probes in the companion model repo. Base model: CohereLabs/tiny-aya-global.
Headline (test split, N=4847, base pos_rate=0.8106044976273984)
| system | AURC | risk @ 80% cov. | risk @ 90% cov. |
|---|---|---|---|
S0_no_guard |
0.811 | 0.811 | 0.811 |
probe_b (S5a) |
0.582 | 0.765 | 0.790 |
probe_a (S5b) |
0.600 | 0.765 | 0.790 |
length |
0.763 | 0.793 | 0.804 |
random |
0.815 | 0.816 | 0.813 |
Plot: eval/test/risk_coverage.png.
SAE alignment (M10 Phase 1)
Probe projected onto the Tiny Aya Global layer-28
SAE (Farseen0/tiny-aya-saes, 16,384 features, BatchTopK).
|cos|_max=0.09690588712692261 — the probe is a distributed
direction, not a single feature. Top features (alignment.json has
full exemplars):
| feature | cos to probe | fired on |
|---|---|---|
| f1358 | +0.097 | 3/14633 |
| f14673 | +0.091 | 40/14633 |
| f10036 | -0.089 | 15/14633 |
| f235 | +0.087 | 3282/14633 |
| f9533 | -0.087 | 52/14633 |
Files
eval/test/metrics.json— per-system AURC, Risk@80/90, SimpleQA penalty, bootstrap CIs, per-dataset breakdowns.eval/test/risk_coverage.png— overlay plot (five systems + S0 base-rate line).interp/sae_alignment/alignment.json— top-k features with cosine, firing counts, max-activating exemplars.interp/sae_alignment/alignment.png— cosine histogram + top-feature bar.docs/FINDINGS.md— the written analysis.
Reproduction
Pair with the companion probe repo. See configs/default.yaml there
for the exact split hashes.
- Downloads last month
- 74