The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
row_id: string
source_dataset: string
source_id: string
question: string
answer: string
context: string
row_type: string
-- schema metadata --
huggingface: '{"info": {"features": {"row_id": {"dtype": "string", "_type' + 324
to
{'row_id': Value('string'), 'question': Value('string'), 'answer': Value('string'), 'context': Value('string'), 'row_type': Value('string'), 'context_words': Value('int64'), 'final_url': Value('string'), 'document_id': Value('string'), 'document_source': Value('string'), 'document_url': Value('string'), 'question_id': Value('string'), 'question_type': Value('string'), 'question_focus': Value('string'), 'umls_cui': Value('string'), 'umls_semantic_group': Value('string'), 'synonyms': 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 2840, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, 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 536, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, 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/parquet/parquet.py", line 220, in _generate_tables
yield Key(file_idx, batch_idx), self._cast_table(pa_table)
~~~~~~~~~~~~~~~~^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/parquet/parquet.py", line 156, 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
row_id: string
source_dataset: string
source_id: string
question: string
answer: string
context: string
row_type: string
-- schema metadata --
huggingface: '{"info": {"features": {"row_id": {"dtype": "string", "_type' + 324
to
{'row_id': Value('string'), 'question': Value('string'), 'answer': Value('string'), 'context': Value('string'), 'row_type': Value('string'), 'context_words': Value('int64'), 'final_url': Value('string'), 'document_id': Value('string'), 'document_source': Value('string'), 'document_url': Value('string'), 'question_id': Value('string'), 'question_type': Value('string'), 'question_focus': Value('string'), 'umls_cui': Value('string'), 'umls_semantic_group': Value('string'), 'synonyms': 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.
Medical hallucination detection
A dataset for training a small model to detect hallucinations in medical answers
and explain why, by checking each answer against the context it should be
grounded in. Each row is a (question, answer, context) triple with a row_type:
not_hallucinated-- the answer is grounded in its context.hallucinated-- the answer is not (sourced separately; see below).
The not_hallucinated split (this build)
Derived from MedQuAD, a collection of medical question-answer pairs curated from 12 US National Institutes of Health websites. MedQuAD ships each answer with the URL of the document it was written from, but no context text. This build:
- Takes MedQuAD rows whose source document is still live and scrapable in 2026 (CancerGov (666), GHR (5285), MPlusHealthTopics (974), NHLBI (539)). GARD, NIDDK and NIHSeniorHealth links are dead; ADAM and MedlinePlus Drugs answers are absent from the source dump.
- Scrapes each source document and extracts its main content as inline
context. - Emits the
(question, answer, context)triple.
Because each answer was authored from its own source document, it is grounded in
that context by construction -- so every row in this split is not_hallucinated.
Full page text is kept as context (no truncation); context_words records length.
The hallucinated split
Sourced separately (e.g. from a hallucination benchmark) and pushed into the same dataset. Note the two splits may differ in register and context format; consumers building a detector should be aware a model can otherwise shortcut on source style.
Fields
Columns are source-agnostic so grounded and hallucinated rows share one schema:
| field | meaning |
|---|---|
question, answer, context |
the training triple |
row_type |
not_hallucinated / hallucinated |
row_id |
globally unique, {source}::{source_id} |
source_dataset |
origin dataset (e.g. lavita/MedQuAD) |
source_id |
identifier within the origin dataset; trace back for any further metadata |
Attribution & licensing
Built on MedQuAD -- Asma Ben Abacha and Dina Demner-Fushman, "A Question-Entailment Approach to Question Answering", BMC Bioinformatics 2019. Source pages are public health content from NIH/NLM. Respect the terms of MedQuAD and of each source site; this dataset redistributes derived context for research use. Please cite MedQuAD.
- Downloads last month
- 40