The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
Error code: DatasetGenerationError
Exception: CastError
Message: Couldn't cast
artifacts: struct<candidate_entities_valid.json: string, eval_targets_valid.json: string, known_true_valid.json (... 9 chars omitted)
child 0, candidate_entities_valid.json: string
child 1, eval_targets_valid.json: string
child 2, known_true_valid.json: string
fold: string
invariants: struct<candidate_partition: string, contains_test_artifact_metadata: bool, known_true_partition: str (... 4 chars omitted)
child 0, candidate_partition: string
child 1, contains_test_artifact_metadata: bool
child 2, known_true_partition: string
partition_hashes: struct<train_entities.json: string, valid_entities.json: string>
child 0, train_entities.json: string
child 1, valid_entities.json: string
protocol: string
runtime_input_hashes: struct<descriptions.json: string, entities.json: string, relation_names.json: string>
child 0, descriptions.json: string
child 1, entities.json: string
child 2, relation_names.json: string
text: int64
to
{'text': Value('int64')}
because column names don't match
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1827, in _prepare_split_single
for key, table in generator:
^^^^^^^^^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
for item in generator(*args, **kwargs):
~~~~~~~~~^^^^^^^^^^^^^^^^^
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
artifacts: struct<candidate_entities_valid.json: string, eval_targets_valid.json: string, known_true_valid.json (... 9 chars omitted)
child 0, candidate_entities_valid.json: string
child 1, eval_targets_valid.json: string
child 2, known_true_valid.json: string
fold: string
invariants: struct<candidate_partition: string, contains_test_artifact_metadata: bool, known_true_partition: str (... 4 chars omitted)
child 0, candidate_partition: string
child 1, contains_test_artifact_metadata: bool
child 2, known_true_partition: string
partition_hashes: struct<train_entities.json: string, valid_entities.json: string>
child 0, train_entities.json: string
child 1, valid_entities.json: string
protocol: string
runtime_input_hashes: struct<descriptions.json: string, entities.json: string, relation_names.json: string>
child 0, descriptions.json: string
child 1, entities.json: string
child 2, relation_names.json: string
text: int64
to
{'text': Value('int64')}
because column names don't match
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
~~~~~~~~~~~~~~~~~~~~~~~~~^
builder, max_dataset_size_bytes=max_dataset_size_bytes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
for job_id, done, content in self._prepare_split_single(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
):
^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1879, in _prepare_split_single
raise DatasetGenerationError("An error occurred while generating the dataset") from e
datasets.exceptions.DatasetGenerationError: An error occurred while generating the datasetNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
text int64 |
|---|
0 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
11 |
13 |
15 |
16 |
17 |
18 |
21 |
22 |
23 |
24 |
25 |
26 |
27 |
28 |
32 |
33 |
34 |
35 |
37 |
38 |
39 |
42 |
43 |
44 |
46 |
47 |
48 |
49 |
50 |
51 |
52 |
53 |
54 |
55 |
56 |
57 |
58 |
59 |
60 |
61 |
62 |
63 |
64 |
65 |
67 |
68 |
69 |
70 |
71 |
73 |
76 |
77 |
78 |
79 |
80 |
81 |
83 |
84 |
85 |
86 |
87 |
89 |
90 |
91 |
92 |
96 |
97 |
98 |
99 |
101 |
103 |
104 |
105 |
106 |
107 |
108 |
110 |
111 |
112 |
114 |
115 |
117 |
118 |
119 |
120 |
121 |
122 |
123 |
125 |
126 |
128 |
DBP5L-Ind
Private research staging. Not a cleared public dataset release or a complete numerical-reproduction bundle.
DBP5L-Ind is a multilingual, concept-disjoint inductive knowledge-graph benchmark candidate covering English, French, Spanish, Japanese and Greek. The browsable files below are byte-identical copies of the 70 members in the privately staged RC1 archive. Only the repository layout and navigation have changed; no scientific records or partition memberships have been edited.
Repository guide
| Folder / file | Contents |
|---|---|
| data/text/ | Frozen entity descriptions, text provenance and text manifest |
| data/identity/ | Identity and co-partition-unit mappings |
| data/folds/ | TRAIN/validation membership and validation-only protocol artifacts |
| attribution/ | Per-record attribution, coverage and component licence maps |
| manifests/ | Original package/transport manifests, source-change metadata and withheld-artifact seals |
| docs/ | Reconstruction notes, modifications, draft licences/notices and original hold markers |
| REPOSITORY_LAYOUT.json | Exact original-archive to repository-path mapping, lengths and hashes |
| FILE_SHA256SUMS.txt | SHA256 for every expanded payload file |
| dataset_RC1_DO_NOT_UPLOAD.zip | Unchanged original transport snapshot, retained for provenance |
Original manifests and documentation describe original archive paths. Consult REPOSITORY_LAYOUT.json when locating files moved into docs/, attribution/ or manifests/. Historical data directory names remain unchanged for compatibility:
| Partition label | Physical directory | Split seed |
|---|---|---|
| P0 | data/folds/fold0_seed13/ |
13 |
| P1 | data/folds/fold1_seed42/ |
42 |
| P2 | data/folds/fold2_seed79/ |
79 |
Dataset contents
- 56,589 entity-text records: 56,083 frozen Wikipedia introduction extracts and 506 label-only fallbacks.
- Text provenance and per-record attribution, identity/co-partition maps, scoped TRAIN/validation membership and validation artifacts.
- Three partitions P0, P1 and P2. Seeds 13, 42 and 79 are unchanged.
Protected test-side files, raw experimental logs/ranks and model weights are excluded. Upstream raw KEnS files and a training-triple export are not included. Withheld evaluation artifacts are represented by seals, not their contents. This repository therefore does not support a claim of complete numerical reproduction.
Integrity
The original archive is 67,419,592 bytes, SHA256:
89ed5e5ddb755d82f85a04a4e009416cf9dc919ef48251eb41fa810216efba28
Its checksum remains in SHA256SUMS.txt. Expanded files are independently listed in FILE_SHA256SUMS.txt. REPOSITORY_LAYOUT.json records their original archive paths and their current repository paths. Layout changes do not close any release gate.
Provenance and release limitations
Frozen text was retrieved on 2026-09-09. The extract service did not take a revision parameter; an observed revision ID is not proof of exact historical extract binding.
Release preparation checked 56,259 distinct observed revisions: 56,258 matched their recorded page IDs and hashes; one unavailable French revision belongs to a label-only record. Category screening inspected 27,042 English/French pages, with one unavailable French label-only page retained as an exception. Spanish, Japanese and Greek were not screened by that category procedure.
33 pages were flagged for imported-content attribution review. Source inspection yielded 54 heuristic notice candidates; dispositions remain unresolved. Current page categories do not establish whether imported material occurs in the frozen introduction. These checks are not universal copyright audit or legal clearance.
Public-release rights, scope, source-notice, independent exact-byte review, human anonymity review, venue-link and publication approvals remain unresolved. All-entity text/maps plus TRAIN/validation membership can expose held-out membership by complement; this scope boundary remains under review. No private authorization/review records or repository history are included in the payload.
Licences and contact
Original source notices and applicable licences remain in effect. CC BY-SA 4.0 is the selected route for eligible Wikipedia-derived adaptations, subject to final per-component review; it is not a blanket grant of rights for all upstream content. Draft licences, notices and HOLD markers are retained without asserting completed public clearance.
Dataset contact: dbp5l.ind.release@gmail.com.
Keep this repository private. Do not share it as a submission-ready released benchmark or make it public until the required approvals are recorded for the exact final package.
- Downloads last month
- 671