Datasets:

frankyifei
/

politnuggets

Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
official_id: string
official_name: string
experiment_type: string
source_columns: list<item: string>
  child 0, item: string
source_count: int64
char_count: int64
token_count: int64
content: string
source_urls: list<item: string>
  child 0, item: string
to
{'official_id': Value('string'), 'official_name': Value('string'), 'experiment_type': Value('string'), 'source_urls': List(Value('string')), 'source_count': Value('int64'), 'char_count': Value('int64'), 'token_count': Value('int64'), 'content': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1779, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 609, in wrapped
                  for item in generator(*args, **kwargs):
                              ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 299, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 128, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2321, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2249, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              official_id: string
              official_name: string
              experiment_type: string
              source_columns: list<item: string>
                child 0, item: string
              source_count: int64
              char_count: int64
              token_count: int64
              content: string
              source_urls: list<item: string>
                child 0, item: string
              to
              {'official_id': Value('string'), 'official_name': Value('string'), 'experiment_type': Value('string'), 'source_urls': List(Value('string')), 'source_count': Value('int64'), 'char_count': Value('int64'), 'token_count': Value('int64'), 'content': Value('string')}
              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 1342, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 907, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1646, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1832, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

official_id string	official_name string	experiment_type string	source_urls list	source_count int64	char_count int64	token_count int64	content string
Administrator_of_the_Environmental_Protection_Agency_Carol_Browner_1993	Carol Browner	unique_related	["http://www.nytimes.com/library/politics/whouse/articles/121792browner-profile.html","https://albri(...TRUNCATED)	73	2,622,817	575,131	"\n\n--- Document Source: http://www.nytimes.com/library/politics/whouse/articles/121792browner-prof(...TRUNCATED)
Administrator_of_the_Environmental_Protection_Agency_Catherine_McCabe_2017	Catherine McCabe	unique_related	["https://19january2017snapshot.epa.gov/aboutepa/catherine-mccabe-deputy-regional-administrator-epas(...TRUNCATED)	63	1,467,543	371,601	"\n\n--- Document Source: https://19january2017snapshot.epa.gov/aboutepa/catherine-mccabe-deputy-reg(...TRUNCATED)
Administrator_of_the_Environmental_Protection_Agency_Gina_McCarthy_2013	Gina McCarthy	unique_related	["http://www.allgov.com/officials/mccarthy-gina?officialid=29790","https://19january2017snapshot.epa(...TRUNCATED)	75	2,044,168	450,528	"\n\n--- Document Source: http://www.allgov.com/officials/mccarthy-gina?officialid=29790 ---\nAllGov(...TRUNCATED)
Administrator_of_the_Small_Business_Administration_A¨ªda_lvarez_1997	A¨ªda _lvarez	unique_related	["http://calhum.org/files/uploads/other/A_State_of_Open_Mind_Magazine_Spring_2013_web.pdf","http://d(...TRUNCATED)	112	5,481,609	1,364,945	"\n\n--- Document Source: http://calhum.org/files/uploads/other/A_State_of_Open_Mind_Magazine_Spring(...TRUNCATED)
Administrator_of_the_Small_Business_Administration_Jeanne_Hulit_2013	Jeanne Hulit	unique_related	["http://lldc.mainelegislature.org/Open/Elections/Returns/2008/20081104/elec_20081104c.pdf","https:/(...TRUNCATED)	101	2,287,234	521,025	"\n\n--- Document Source: http://lldc.mainelegislature.org/Open/Elections/Returns/2008/20081104/elec(...TRUNCATED)
Administrator_of_the_Small_Business_Administration_Jovita_Carranza_2020	Jovita Carranza	unique_related	["http://media.corporate-ir.net/media_files/NYS/UPS/reports/ar2000/pdf/UPS2000ARFINAL.pdf","https://(...TRUNCATED)	81	3,396,273	795,944	"\n\n--- Document Source: http://media.corporate-ir.net/media_files/NYS/UPS/reports/ar2000/pdf/UPS20(...TRUNCATED)
Administrator_of_the_Small_Business_Administration_Linda_McMahon_2017	Linda McMahon	unique_related	["http://www.allgov.com/departments/independent-agencies/small-business-administration?agencyid=7308(...TRUNCATED)	99	2,830,711	630,315	"\n\n--- Document Source: http://www.allgov.com/departments/independent-agencies/small-business-admi(...TRUNCATED)
Administrator_of_the_Small_Business_Administration_Marianne_Markowitz_2014	Marianne Markowitz	unique_related	["http://apps.chicagotribune.com/ebooks/innovationAndTechnology.pdf?track=bluesky-newsletter-weekly"(...TRUNCATED)	75	2,970,105	632,942	"\n\n--- Document Source: http://apps.chicagotribune.com/ebooks/innovationAndTechnology.pdf?track=bl(...TRUNCATED)
Administrator_of_the_Small_Business_Administration_Sandy_Baruah_2008	Sandy Baruah	unique_related	["http://www.allgov.com/departments/department-of-commerce/economic-development-administration?agenc(...TRUNCATED)	78	2,121,893	486,827	"\n\n--- Document Source: http://www.allgov.com/departments/department-of-commerce/economic-developm(...TRUNCATED)
Ambassador_To_The_Us_Bozo_Cerar_2014	Bozo Cerar	unique_related	["http://countrylicious.com/slovenia","http://www.allgov.com/news/appointments-and-resignation/slove(...TRUNCATED)	52	1,967,178	613,636	"\n\n--- Document Source: http://countrylicious.com/slovenia ---\nSlovenia country facts\n\n========(...TRUNCATED)

End of preview.

PolitNuggets

Replication artifacts for our paper on LLM-based extraction of political-elite biographies. The companion source code for the deep-research API lives in a separate GitHub repository (poli_searcher).

This dataset contains three data artifacts plus a descriptive figure:

1. Consolidated Ground Truth (`cgt/cgt.csv`)

The primary same-schema CGT table for all 398 entities. It is produced by the main CGT synthesis pipeline, which deduplicates near-duplicate entries, prefers more complete facts, and uses majority consensus when source variants conflict. This file is intentionally narrower than the union-style search_master_cgt_comprehensive_en.csv.

Columns: official_id, cgt_party_en, cgt_occupation_en, cgt_education_en, cgt_relatives_en, plus generation status/model metadata.

2. Cached webpages (`cached_webpages/`)

Raw retrieved pages collected during our agentic runs. These fix the search snapshot used for every reported number and enable offline re-evaluation.

cached_webpages/raw_doc_store.sql - SQLite dump consumed by langchain_community.storage.SQLStore (namespace raw_documents_ns).
cached_webpages/chunk_registry.json - URL-to-storage-key registry.
cached_webpages/chroma.sqlite3, record_manager_cache.sql - Chroma metadata / LangChain indexing cache.

Open the store in Python::

from langchain_community.storage import SQLStore
store = SQLStore(
    namespace="raw_documents_ns",
    db_url="sqlite:///cached_webpages/raw_doc_store.sql",
)
bytes_ = store.mget(["<chunk-key>"])[0]

3. LRM evaluation package (`lrm_eval/`)

Curated static-context dataset for evaluating long-context biography extraction without interactive search, for controlled comparison of "Reasoning in Context" across models.

lrm_eval/archive_short/unique_related.jsonl - related-URL aggregated text.
lrm_eval/archive_short/unique_retrieved.jsonl - retrieved-URL aggregated text.
lrm_eval/archive_short/unique_wiki.jsonl - wiki/encyclopedia aggregated text.
lrm_eval/long_context/lc_sup_plus.jsonl - supervisor + detailed markdown report built by our grok_v7 agent (text-only).
lrm_eval/long_context/lc_retrieved.jsonl - retrieved-URL aggregated text for the long-context URL setting.

Each JSONL row has::

{
  "official_id": ...,
  "official_name": ...,
  "experiment_type": ...,
  "source_urls"|"source_columns": [...],
  "source_count": ...,
  "char_count": ...,
  "token_count": ...,
  "content": "..."
}

Descriptive figure: language share by country

The dataset card includes a country-level language-share figure comparing the non-English token share in retrieved long-context evidence with the non-English token share after the synthesized/refined long-context construction.

Companion code

Deep-research API (poli_searcher): .

Citation

{{ bibtex entry here }}

Downloads last month: 76

Total file size:

9.15 GB