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
annotator: string
gold: struct<2075987c76c3922b: struct<52.203-11: string, 52.203-2: string, 52.204-12: string, 52.204-25: s (... 9952 chars omitted)
child 0, 2075987c76c3922b: struct<52.203-11: string, 52.203-2: string, 52.204-12: string, 52.204-25: string, 52.204-26: string, (... 74 chars omitted)
child 0, 52.203-11: string
child 1, 52.203-2: string
child 2, 52.204-12: string
child 3, 52.204-25: string
child 4, 52.204-26: string
child 5, 52.225-1: string
child 6, 52.227-14: string
child 7, 52.227-16: string
child 8, 52.230-1: string
child 1, 2b5c379382e8c098: struct<1452.204-70: string, 1452.228-70: string, 1452.280-2: string, 1452.280-4: string, 52.212-1: s (... 135 chars omitted)
child 0, 1452.204-70: string
child 1, 1452.228-70: string
child 2, 1452.280-2: string
child 3, 1452.280-4: string
child 4, 52.212-1: string
child 5, 52.212-3: string
child 6, 52.212-4: string
child 7, 52.212-5: string
child 8, 52.223-18: string
child 9, 52.225-1: string
child 10, 52.232-33: string
child 11, 52.232-99: string
child 2, 2e6e745d21f6d260: struct<252.204-7012: string, 252.204-7016: string, 252.204-7018: string, 252.204-7019: string, 252.2 (... 322 chars omitted)
child 0, 252.204-7012: string
child 1, 252.204-7016: string
child 2, 252.204-7018: string
child 3, 252.204-7019: string
child 4, 252.204-7020: string
child 5, 2
...
cy: doub (... 33 chars omitted)
child 0, precision: double
child 1, recall: double
child 2, f: double
child 3, f2: double
child 4, specificity: double
child 5, accuracy: double
child 6, f1_lo: double
child 7, f1_hi: double
model max q>=0.5: struct<precision: double, recall: double, f: double, f2: double, specificity: double, accuracy: doub (... 33 chars omitted)
child 0, precision: double
child 1, recall: double
child 2, f: double
child 3, f2: double
child 4, specificity: double
child 5, accuracy: double
child 6, f1_lo: double
child 7, f1_hi: double
B0 (VETR): struct<precision: double, recall: double, f: double, f2: double, specificity: double, accuracy: doub (... 33 chars omitted)
child 0, precision: double
child 1, recall: double
child 2, f: double
child 3, f2: double
child 4, specificity: double
child 5, accuracy: double
child 6, f1_lo: double
child 7, f1_hi: double
_meta: struct<documents: int64, labelled: int64, binding: int64, gold: string, mode: string>
child 0, documents: int64
child 1, labelled: int64
child 2, binding: int64
child 3, gold: string
child 4, mode: string
model max q>=0.9: struct<precision: double, recall: double, f: double, f2: double, specificity: double, accuracy: doub (... 33 chars omitted)
child 0, precision: double
child 1, recall: double
child 2, f: double
child 3, f2: double
child 4, specificity: double
child 5, accuracy: double
child 6, f1_lo: double
child 7, f1_hi: double
to
{'all-candidates': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'B0 (VETR)': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'B0 minus empty-box clauses': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'B1 rules': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model noisy_or q>=0.5': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model noisy_or q>=0.7': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model noisy_or q>=0.9': {
...
2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model max q>=0.5': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model max q>=0.7': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model max q>=0.9': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'paired model noisy_or q>=0.5 vs B1 rules': {'diff': Value('float64'), 'lo': Value('float64'), 'hi': Value('float64'), 'share_not_better': Value('float64')}, 'paired model noisy_or q>=0.5 vs B0 (VETR)': {'diff': Value('float64'), 'lo': Value('float64'), 'hi': Value('float64'), 'share_not_better': Value('float64')}, 'paired model noisy_or q>=0.5 vs all-candidates': {'diff': Value('float64'), 'lo': Value('float64'), 'hi': Value('float64'), 'share_not_better': Value('float64')}, '_meta': {'documents': Value('int64'), 'labelled': Value('int64'), 'binding': Value('int64'), 'gold': Value('string'), 'mode': Value('string')}}
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 2951, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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
annotator: string
gold: struct<2075987c76c3922b: struct<52.203-11: string, 52.203-2: string, 52.204-12: string, 52.204-25: s (... 9952 chars omitted)
child 0, 2075987c76c3922b: struct<52.203-11: string, 52.203-2: string, 52.204-12: string, 52.204-25: string, 52.204-26: string, (... 74 chars omitted)
child 0, 52.203-11: string
child 1, 52.203-2: string
child 2, 52.204-12: string
child 3, 52.204-25: string
child 4, 52.204-26: string
child 5, 52.225-1: string
child 6, 52.227-14: string
child 7, 52.227-16: string
child 8, 52.230-1: string
child 1, 2b5c379382e8c098: struct<1452.204-70: string, 1452.228-70: string, 1452.280-2: string, 1452.280-4: string, 52.212-1: s (... 135 chars omitted)
child 0, 1452.204-70: string
child 1, 1452.228-70: string
child 2, 1452.280-2: string
child 3, 1452.280-4: string
child 4, 52.212-1: string
child 5, 52.212-3: string
child 6, 52.212-4: string
child 7, 52.212-5: string
child 8, 52.223-18: string
child 9, 52.225-1: string
child 10, 52.232-33: string
child 11, 52.232-99: string
child 2, 2e6e745d21f6d260: struct<252.204-7012: string, 252.204-7016: string, 252.204-7018: string, 252.204-7019: string, 252.2 (... 322 chars omitted)
child 0, 252.204-7012: string
child 1, 252.204-7016: string
child 2, 252.204-7018: string
child 3, 252.204-7019: string
child 4, 252.204-7020: string
child 5, 2
...
cy: doub (... 33 chars omitted)
child 0, precision: double
child 1, recall: double
child 2, f: double
child 3, f2: double
child 4, specificity: double
child 5, accuracy: double
child 6, f1_lo: double
child 7, f1_hi: double
model max q>=0.5: struct<precision: double, recall: double, f: double, f2: double, specificity: double, accuracy: doub (... 33 chars omitted)
child 0, precision: double
child 1, recall: double
child 2, f: double
child 3, f2: double
child 4, specificity: double
child 5, accuracy: double
child 6, f1_lo: double
child 7, f1_hi: double
B0 (VETR): struct<precision: double, recall: double, f: double, f2: double, specificity: double, accuracy: doub (... 33 chars omitted)
child 0, precision: double
child 1, recall: double
child 2, f: double
child 3, f2: double
child 4, specificity: double
child 5, accuracy: double
child 6, f1_lo: double
child 7, f1_hi: double
_meta: struct<documents: int64, labelled: int64, binding: int64, gold: string, mode: string>
child 0, documents: int64
child 1, labelled: int64
child 2, binding: int64
child 3, gold: string
child 4, mode: string
model max q>=0.9: struct<precision: double, recall: double, f: double, f2: double, specificity: double, accuracy: doub (... 33 chars omitted)
child 0, precision: double
child 1, recall: double
child 2, f: double
child 3, f2: double
child 4, specificity: double
child 5, accuracy: double
child 6, f1_lo: double
child 7, f1_hi: double
to
{'all-candidates': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'B0 (VETR)': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'B0 minus empty-box clauses': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'B1 rules': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model noisy_or q>=0.5': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model noisy_or q>=0.7': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model noisy_or q>=0.9': {
...
2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model max q>=0.5': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model max q>=0.7': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'model max q>=0.9': {'precision': Value('float64'), 'recall': Value('float64'), 'f': Value('float64'), 'f2': Value('float64'), 'specificity': Value('float64'), 'accuracy': Value('float64'), 'f1_lo': Value('float64'), 'f1_hi': Value('float64')}, 'paired model noisy_or q>=0.5 vs B1 rules': {'diff': Value('float64'), 'lo': Value('float64'), 'hi': Value('float64'), 'share_not_better': Value('float64')}, 'paired model noisy_or q>=0.5 vs B0 (VETR)': {'diff': Value('float64'), 'lo': Value('float64'), 'hi': Value('float64'), 'share_not_better': Value('float64')}, 'paired model noisy_or q>=0.5 vs all-candidates': {'diff': Value('float64'), 'lo': Value('float64'), 'hi': Value('float64'), 'share_not_better': Value('float64')}, '_meta': {'documents': Value('int64'), 'labelled': Value('int64'), 'binding': Value('int64'), 'gold': Value('string'), 'mode': 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.
fedproc-ledger data (research release, v1.0)
mentions.parquet: 685,079 clause-number mentions from 6,606 public SAM.gov solicitation documents with status-quo entries: doc_id, number, alternate, role, binding_prob, source (model,box_rule,para_a_ruleortext_rule; rules v1.2, so it differs from the round-1 table in its box and rule counts), confidence, and a 300-character context snippet (personal-data patterns redacted at extraction). Roles and probabilities are model output, not labels.results/: gold label files (ledger_gold_*.json: development set, frozen test and rounds 2 to 7 including the agent, judge and majority files; labels BINDS/NOT/REFERENCED/UNDECIDED from the coding agent and two OpenAI models, no human annotators), evaluation outputs and the frozen document lists with sha256. Round 7 (27 documents posted 2026-09-30 to 2026-10-07, rules v1.4) is the temporal hold-out; the per-mention table refresh with v1.4 predictions is queued, somentions.parquetstays on rules v1.2 for now. Source documents are public US federal solicitation notices; this release contains derived tables and short snippets only (contact emails and phone numbers redacted, hardened against glued tokens in D-045). Released publicly by the CTO of Acu-Elligent LLC for research reproducibility alongside the paper.
- Downloads last month
- -