Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code: DatasetGenerationError
Exception: TypeError
Message: Couldn't cast array of type
struct<id: string, source: string, variant: string, group_id: string, split: string, license: string, label_source: string, provenance: string, main_cpc: string, filing_date: string, layout: string, phrasing: string, sic: string, year: int64, words: int64>
to
{'id': Value('string'), 'source': Value('string'), 'variant': Value('string'), 'group_id': Value('string'), 'split': Value('string'), 'license': Value('string'), 'label_source': Value('string'), 'provenance': Value('string'), 'sic': Value('string'), 'year': Value('int64'), 'words': Value('int64'), 'layout': Value('string'), 'phrasing': Json(decode=True)}
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1835, in _prepare_split_single
for key, table in generator:
^^^^^^^^^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 621, in wrapped
for item in generator(*args, **kwargs):
~~~~~~~~~^^^^^^^^^^^^^^^^^
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 2315, in cast_table_to_schema
cast_array_to_feature(
~~~~~~~~~~~~~~~~~~~~~^
table[name] if name in table_column_names else pa.array([None] * len(table), type=schema.field(name).type),
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
feature,
^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1861, in wrapper
return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
~~~~^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2161, in cast_array_to_feature
raise TypeError(f"Couldn't cast array of type\n{_short_str(array.type)}\nto\n{_short_str(feature)}")
TypeError: Couldn't cast array of type
struct<id: string, source: string, variant: string, group_id: string, split: string, license: string, label_source: string, provenance: string, main_cpc: string, filing_date: string, layout: string, phrasing: string, sic: string, year: int64, words: int64>
to
{'id': Value('string'), 'source': Value('string'), 'variant': Value('string'), 'group_id': Value('string'), 'split': Value('string'), 'license': Value('string'), 'label_source': Value('string'), 'provenance': Value('string'), 'sic': Value('string'), 'year': Value('int64'), 'words': Value('int64'), 'layout': Value('string'), 'phrasing': Json(decode=True)}
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 1379, 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 956, 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 1702, 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 1887, 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.
state unknown | questions unknown | _meta dict |
|---|---|---|
{
"form": "10-K",
"filer": "META GROUP INC",
"section": "Item 1. Business",
"text": "ITEM 1. BUSINESS GENERAL META Group, Inc., a Delaware corporation founded in 1989 (\"META Group,\" together with its subsidiaries-collectively the \"Company\"), is a leading provider of information technology (\"IT\") research,... | {
"industry_classification": {
"type": "choice",
"instructions": "Based on this description of the business, which industry does the company mainly operate in?",
"criteria": {
"membership_organizations": null,
"business_services": null,
"transportation_services": null,
"wholesale_t... | {
"id": "d4:10k/1000015_2003",
"source": "d4:10k_sic",
"variant": "clean",
"group_id": "10k/1000015",
"split": "train",
"license": "public domain (SEC EDGAR filing; Item 1 via EDGAR-CORPUS, Apache-2.0)",
"label_source": "publisher: the filer's SIC code in SEC's submissions data (current code)",
"provena... |
{
"form": "10-K",
"filer": "META GROUP INC",
"section": "Item 1. Business",
"text": "ITEM 1. BUSINESS General META Group, Inc. (\"META Group\") and its subsidiary, The Sentry Group, Inc. (\"Sentry\"), collectively, the \"Company\" is an independent market assessment company providing research and analysis of de... | {
"sector": {
"type": "choice",
"instructions": "In which economic sector does this registrant chiefly operate?",
"criteria": {
"agriculture_forestry_fishing": null,
"mining": null,
"construction": null,
"manufacturing": null,
"transportation_and_utilities": null,
"whol... | {
"id": "d4:10k/1000015_1998",
"source": "d4:10k_sic",
"variant": "clean",
"group_id": "10k/1000015",
"split": "train",
"license": "public domain (SEC EDGAR filing; Item 1 via EDGAR-CORPUS, Apache-2.0)",
"label_source": "publisher: the filer's SIC code in SEC's submissions data (current code)",
"provena... |
"Item 1. Business General Nicholas Financial, Inc. (“Nicholas Financial-Canada”) is a Canadian holding company incorporated under the laws of British Columbia in 1986. The business activities of Nicholas Financial-Canada are currently conducted exclusively through its wholly-owned indirect subsidiary, Nicholas Financia... | {
"major_group": {
"type": "choice",
"instructions": "Based on this description of the business, which industry does the company mainly operate in?",
"criteria": {
"10": "metal mining — includes: gold and silver ores; miscellaneous metal ores",
"12": "bituminous coal & lignite mining; bitumino... | {
"id": "d4:10k/1000045_2018",
"source": "d4:10k_sic",
"variant": "clean",
"group_id": "10k/1000045",
"split": "train",
"license": "public domain (SEC EDGAR filing; Item 1 via EDGAR-CORPUS, Apache-2.0)",
"label_source": "publisher: the filer's SIC code in SEC's submissions data (current code)",
"provena... |
"Item 1. Business General Nicholas Financial, Inc. (“Nicholas Financial-Canada”) is a Canadian holding company incorporated under the laws of British Columbia in 1986. The business activities of Nicholas Financial-Canada are conducted through its two wholly-owned subsidiaries formed pursuant to the laws of the State of... | {
"industry": {
"type": "choice",
"instructions": "Going by what the company says it does, which of these industries is its main line of business?",
"criteria": {
"10": "Metal Mining",
"12": "Coal Mining",
"13": "Oil and Gas Extraction",
"14": "Mining and Quarrying of Nonmetallic M... | {
"id": "d4:10k/1000045_2012",
"source": "d4:10k_sic",
"variant": "clean",
"group_id": "10k/1000045",
"split": "train",
"license": "public domain (SEC EDGAR filing; Item 1 via EDGAR-CORPUS, Apache-2.0)",
"label_source": "publisher: the filer's SIC code in SEC's submissions data (current code)",
"provena... |
"ITEM 1. BUSINESS. - ----------------- GENERAL Lenox Bancorp, Inc. (also referred to as the \"Compan(...TRUNCATED) | {"broad_sector":{"type":"choice","instructions":"Judging by the business description, which division(...TRUNCATED) | {"id":"d4:10k/1000050_1996","source":"d4:10k_sic","variant":"clean","group_id":"10k/1000050","split"(...TRUNCATED) |
"ITEM 1 - BUSINESS GENERAL 5B Technologies Corporation (formerly Paramount Financial Corporation) an(...TRUNCATED) | {"economic_sector":{"type":"choice","instructions":"In which economic sector does this registrant ch(...TRUNCATED) | {"id":"d4:10k/1000179_2000","source":"d4:10k_sic","variant":"clean","group_id":"10k/1000179","split"(...TRUNCATED) |
"Item 1. Business On August 12, 1996, Registrant, through the efforts of its General Partner, commen(...TRUNCATED) | {"industry_classification":{"type":"choice","instructions":"Based on this description of the busines(...TRUNCATED) | {"id":"d4:10k/1000181_2000","source":"d4:10k_sic","variant":"clean","group_id":"10k/1000181","split"(...TRUNCATED) |
{"form":"10-K","filer":"STERLING HOUSE CORP","section":"Item 1. Business","text":"ITEM 1. Business G(...TRUNCATED) | {"sic_major_group":{"type":"choice","instructions":"Which SIC major group best describes this compan(...TRUNCATED) | {"id":"d4:10k/1000185_1996","source":"d4:10k_sic","variant":"clean","group_id":"10k/1000185","split"(...TRUNCATED) |
{"form":"10-K","filer":"ENTERPRISE SYSTEMS INC /DE/","section":"Item 1. Business","text":"Item 1. Bu(...TRUNCATED) | {"major_group":{"type":"choice","instructions":"Which SIC major group best describes this company's (...TRUNCATED) | {"id":"d4:10k/1000188_1996","source":"d4:10k_sic","variant":"clean","group_id":"10k/1000188","split"(...TRUNCATED) |
{"form":"10-K","filer":"OCAL INC","section":"Item 1. Business","text":"ITEM 1. BUSINESS. GENERAL ___(...TRUNCATED) | {"line_of_business":{"type":"choice","instructions":"Going by what the company says it does, which o(...TRUNCATED) | {"id":"d4:10k/1000189_1997","source":"d4:10k_sic","variant":"clean","group_id":"10k/1000189","split"(...TRUNCATED) |
End of preview.
encoder-rtx D4: labelled public records
Long documents with publisher labels, as decision records for the encoder-rtx decision stage
(stacks, stacks/encoder-rtx/): Item 1 (Business)
of SEC 10-K filings labelled with the filer's SIC code, and US patent applications labelled
with their CPC codes. Each record is one document plus 1-4 questions about it in Kev's suite
schema (state, questions with type / instructions / criteria / label, _meta):
- 10-Ks: the SIC division (choice over 10), the 2-digit major group (choice: TypeSafe's 75-group option set, the SIC manual's 72 named groups, or a 8-24 option subset), the 3-digit industry group within the gold major group, and yes/no questions on a division or group.
- Patents: the CPC section (choice over A-H), the class within the gold section, the subclass within the gold class, and yes/no questions on a subclass (any of the application's CPC labels) or a section (the main label's).
Yes/no questions alternate the asked class per template, so each template is exactly half true. Every question type has several phrasings, and the states come in several layouts.
Splits (by company / application)
| file | what |
|---|---|
data/train.jsonl |
94% of companies / applications; the train phrasings and layouts |
data/dev.jsonl |
3%; the same phrasings (in-distribution) |
data/holdout.jsonl |
3%; phrasings and a state layout used nowhere else (does the task transfer off its templates?) |
raw/ |
the inputs as fetched (code/fetch_d4.py) |
stats.json, screen_report.json |
counts, option sets, token lengths, yes/no balance; the overlap screen |
Every record was screened against the decision stage's evaluation partitions (Kev's suites, JevBench's public items, Recast NLP's holdout, Cookbook-dev) with Kev's overlap rule; the 60 companies in Cookbook-dev's 10-K set are left out entirely.
Sources and licenses
- 10-K text: SEC EDGAR filings (US government public records, public domain), Item 1 as
extracted by EDGAR-CORPUS
(Loukas et al., 2021; Apache-2.0) @
7e90f0f3, validation and test files 1993-2020. - SIC codes and company names: SEC's bulk
submissions.zip(public domain), the code each filer carries today. A company that changed industry since an older filing carries its current code. - SIC titles: the SIC manual's division and major-group titles (US Department of Labor, public domain); TypeSafe's 75-group option text (MIT, from its public cookbook repo); SEC's 4-digit code titles.
- Patents: HUPD (Suzgun et al., 2022;
CC BY-SA 4.0) @
f570a84b, the January 2016 sample and the 2018 year: title, abstract, claims and CPC labels. - CPC titles: the Cooperative Patent Classification title list (EPO / USPTO), 2026.08.
The compilation is released under CC BY-SA 4.0, as HUPD's share-alike license requires.
Built by agent-ops-stacks/encoder-rtx/2026-11-01_0833am_d4-public-records/
(fetch_d4.py, render_d4.py, build_d4.py).
- Downloads last month
- 144