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
model: string
N_per_family: int64
cap: int64
n_pairs: int64
mixed_items_per_family: struct<speaker: int64, literal: int64, presup: int64>
child 0, speaker: int64
child 1, literal: int64
child 2, presup: int64
pair_pool_per_family: struct<speaker: int64, literal: int64, presup: int64>
child 0, speaker: int64
child 1, literal: int64
child 2, presup: int64
sets: struct<folio: int64, gsm8k: int64, imppres_truth: int64, imppres_presup: int64, pub_task11: int64, p (... 102 chars omitted)
child 0, folio: int64
child 1, gsm8k: int64
child 2, imppres_truth: int64
child 3, imppres_presup: int64
child 4, pub_task11: int64
child 5, pub_task12: int64
child 6, circa: int64
child 7, impl1k: int64
child 8, imppres_fel: int64
child 9, pub_task4: int64
child 10, pub_task5: int64
mean_think_tokens_chosen: double
mean_think_tokens_rejected: double
share_rejected_longer: double
shuffled_share_flipped: double
selected: struct<n: int64, mean_think_tokens: double, share_correct: double, total_tokens: int64, by_family_me (... 68 chars omitted)
child 0, n: int64
child 1, mean_think_tokens: double
child 2, share_correct: double
child 3, total_tokens: int64
child 4, by_family_mean_tokens: struct<speaker: double, literal: double, presup: double>
child 0, speaker: double
child 1, literal: double
child 2, presup: double
control: struct<n: int64, mean_think_tokens: double, share_correct: double, total_tokens: int64, by_family_me (... 68 chars omi
...
int64, zero: int64, pass_k_histogram: struct<4/4: int64, 3/ (... 165 chars omitted)
child 0, items: int64
child 1, mixed: int64
child 2, all4: int64
child 3, zero: int64
child 4, pass_k_histogram: struct<4/4: int64, 3/4: int64, 0/4: int64, 2/4: int64, 1/4: int64>
child 0, 4/4: int64
child 1, 3/4: int64
child 2, 0/4: int64
child 3, 2/4: int64
child 4, 1/4: int64
child 5, selected: int64
child 6, selected_share_4of4: double
child 7, selected_sets: struct<imppres_truth: int64, gsm8k: int64, folio: int64>
child 0, imppres_truth: int64
child 1, gsm8k: int64
child 2, folio: int64
child 2, presup: struct<items: int64, mixed: int64, all4: int64, zero: int64, pass_k_histogram: struct<3/4: int64, 0/ (... 176 chars omitted)
child 0, items: int64
child 1, mixed: int64
child 2, all4: int64
child 3, zero: int64
child 4, pass_k_histogram: struct<3/4: int64, 0/4: int64, 4/4: int64, 1/4: int64, 2/4: int64>
child 0, 3/4: int64
child 1, 0/4: int64
child 2, 4/4: int64
child 3, 1/4: int64
child 4, 2/4: int64
child 5, selected: int64
child 6, selected_share_4of4: double
child 7, selected_sets: struct<pub_task11: int64, pub_task12: int64, imppres_presup: int64>
child 0, pub_task11: int64
child 1, pub_task12: int64
child 2, imppres_presup: int64
to
{'model': Value('string'), 'quota_N_per_family': Value('int64'), 'cap': Value('int64'), 'families': {'speaker': {'items': Value('int64'), 'mixed': Value('int64'), 'all4': Value('int64'), 'zero': Value('int64'), 'pass_k_histogram': {'2/4': Value('int64'), '3/4': Value('int64'), '4/4': Value('int64'), '0/4': Value('int64'), '1/4': Value('int64')}, 'selected': Value('int64'), 'selected_share_4of4': Value('float64'), 'selected_sets': {'circa': Value('int64'), 'imppres_fel': Value('int64'), 'impl1k': Value('int64'), 'pub_task4': Value('int64'), 'pub_task5': Value('int64')}}, 'literal': {'items': Value('int64'), 'mixed': Value('int64'), 'all4': Value('int64'), 'zero': Value('int64'), 'pass_k_histogram': {'4/4': Value('int64'), '3/4': Value('int64'), '0/4': Value('int64'), '2/4': Value('int64'), '1/4': Value('int64')}, 'selected': Value('int64'), 'selected_share_4of4': Value('float64'), 'selected_sets': {'imppres_truth': Value('int64'), 'gsm8k': Value('int64'), 'folio': Value('int64')}}, 'presup': {'items': Value('int64'), 'mixed': Value('int64'), 'all4': Value('int64'), 'zero': Value('int64'), 'pass_k_histogram': {'3/4': Value('int64'), '0/4': Value('int64'), '4/4': Value('int64'), '1/4': Value('int64'), '2/4': Value('int64')}, 'selected': Value('int64'), 'selected_share_4of4': Value('float64'), 'selected_sets': {'pub_task11': Value('int64'), 'pub_task12': Value('int64'), 'imppres_presup': Value('int64')}}}, 'selected': {'n': Value('int64'), 'mean_think_tokens': Value('float64'), 'share_correct': Value('float64'), 'total_tokens': Value('int64'), 'by_family_mean_tokens': {'speaker': Value('float64'), 'literal': Value('float64'), 'presup': Value('float64')}}, 'control': {'n': Value('int64'), 'mean_think_tokens': Value('float64'), 'share_correct': Value('float64'), 'total_tokens': Value('int64'), 'by_family_mean_tokens': {'speaker': Value('float64'), 'literal': Value('float64'), 'presup': Value('float64')}}}
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
model: string
N_per_family: int64
cap: int64
n_pairs: int64
mixed_items_per_family: struct<speaker: int64, literal: int64, presup: int64>
child 0, speaker: int64
child 1, literal: int64
child 2, presup: int64
pair_pool_per_family: struct<speaker: int64, literal: int64, presup: int64>
child 0, speaker: int64
child 1, literal: int64
child 2, presup: int64
sets: struct<folio: int64, gsm8k: int64, imppres_truth: int64, imppres_presup: int64, pub_task11: int64, p (... 102 chars omitted)
child 0, folio: int64
child 1, gsm8k: int64
child 2, imppres_truth: int64
child 3, imppres_presup: int64
child 4, pub_task11: int64
child 5, pub_task12: int64
child 6, circa: int64
child 7, impl1k: int64
child 8, imppres_fel: int64
child 9, pub_task4: int64
child 10, pub_task5: int64
mean_think_tokens_chosen: double
mean_think_tokens_rejected: double
share_rejected_longer: double
shuffled_share_flipped: double
selected: struct<n: int64, mean_think_tokens: double, share_correct: double, total_tokens: int64, by_family_me (... 68 chars omitted)
child 0, n: int64
child 1, mean_think_tokens: double
child 2, share_correct: double
child 3, total_tokens: int64
child 4, by_family_mean_tokens: struct<speaker: double, literal: double, presup: double>
child 0, speaker: double
child 1, literal: double
child 2, presup: double
control: struct<n: int64, mean_think_tokens: double, share_correct: double, total_tokens: int64, by_family_me (... 68 chars omi
...
int64, zero: int64, pass_k_histogram: struct<4/4: int64, 3/ (... 165 chars omitted)
child 0, items: int64
child 1, mixed: int64
child 2, all4: int64
child 3, zero: int64
child 4, pass_k_histogram: struct<4/4: int64, 3/4: int64, 0/4: int64, 2/4: int64, 1/4: int64>
child 0, 4/4: int64
child 1, 3/4: int64
child 2, 0/4: int64
child 3, 2/4: int64
child 4, 1/4: int64
child 5, selected: int64
child 6, selected_share_4of4: double
child 7, selected_sets: struct<imppres_truth: int64, gsm8k: int64, folio: int64>
child 0, imppres_truth: int64
child 1, gsm8k: int64
child 2, folio: int64
child 2, presup: struct<items: int64, mixed: int64, all4: int64, zero: int64, pass_k_histogram: struct<3/4: int64, 0/ (... 176 chars omitted)
child 0, items: int64
child 1, mixed: int64
child 2, all4: int64
child 3, zero: int64
child 4, pass_k_histogram: struct<3/4: int64, 0/4: int64, 4/4: int64, 1/4: int64, 2/4: int64>
child 0, 3/4: int64
child 1, 0/4: int64
child 2, 4/4: int64
child 3, 1/4: int64
child 4, 2/4: int64
child 5, selected: int64
child 6, selected_share_4of4: double
child 7, selected_sets: struct<pub_task11: int64, pub_task12: int64, imppres_presup: int64>
child 0, pub_task11: int64
child 1, pub_task12: int64
child 2, imppres_presup: int64
to
{'model': Value('string'), 'quota_N_per_family': Value('int64'), 'cap': Value('int64'), 'families': {'speaker': {'items': Value('int64'), 'mixed': Value('int64'), 'all4': Value('int64'), 'zero': Value('int64'), 'pass_k_histogram': {'2/4': Value('int64'), '3/4': Value('int64'), '4/4': Value('int64'), '0/4': Value('int64'), '1/4': Value('int64')}, 'selected': Value('int64'), 'selected_share_4of4': Value('float64'), 'selected_sets': {'circa': Value('int64'), 'imppres_fel': Value('int64'), 'impl1k': Value('int64'), 'pub_task4': Value('int64'), 'pub_task5': Value('int64')}}, 'literal': {'items': Value('int64'), 'mixed': Value('int64'), 'all4': Value('int64'), 'zero': Value('int64'), 'pass_k_histogram': {'4/4': Value('int64'), '3/4': Value('int64'), '0/4': Value('int64'), '2/4': Value('int64'), '1/4': Value('int64')}, 'selected': Value('int64'), 'selected_share_4of4': Value('float64'), 'selected_sets': {'imppres_truth': Value('int64'), 'gsm8k': Value('int64'), 'folio': Value('int64')}}, 'presup': {'items': Value('int64'), 'mixed': Value('int64'), 'all4': Value('int64'), 'zero': Value('int64'), 'pass_k_histogram': {'3/4': Value('int64'), '0/4': Value('int64'), '4/4': Value('int64'), '1/4': Value('int64'), '2/4': Value('int64')}, 'selected': Value('int64'), 'selected_share_4of4': Value('float64'), 'selected_sets': {'pub_task11': Value('int64'), 'pub_task12': Value('int64'), 'imppres_presup': Value('int64')}}}, 'selected': {'n': Value('int64'), 'mean_think_tokens': Value('float64'), 'share_correct': Value('float64'), 'total_tokens': Value('int64'), 'by_family_mean_tokens': {'speaker': Value('float64'), 'literal': Value('float64'), 'presup': Value('float64')}}, 'control': {'n': Value('int64'), 'mean_think_tokens': Value('float64'), 'share_correct': Value('float64'), 'total_tokens': Value('int64'), 'by_family_mean_tokens': {'speaker': Value('float64'), 'literal': Value('float64'), 'presup': Value('float64')}}}
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.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Data
new_items/ holds the items created for this paper, released under CC BY 4.0:
- the speaker-knowledge scenarios (
a1ctx_*); - the Bott–Noveck category sentences (
bn_underinformative*); - the context-reversal pairs (
method_ctxrev).
All other item sets are built from their original public releases by code/data_builders/build_items.py, which downloads each source and writes data/items/*.jsonl. They are not redistributed here; use them under their own licences:
| Source | Licence |
|---|---|
| IMPPRES | CC BY-NC 4.0 |
| PUB, Fig-QA, ImplicatureX | MIT |
| LUDWIG, Circa | CC BY-SA 4.0 (Circa) / original terms (LUDWIG) |
| PragMega (Hu et al. 2023) | original terms |
| scalar-diversity sets | as released by the authors |
| NOPE | no licence stated: build locally, do not redistribute |
| FOLIO, GSM8K | CC BY-SA 4.0 / MIT |
See SOURCES_PROVENANCE.md for exact versions.
- Downloads last month
- 493