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
$schema: string
title: string
type: string
required: list<item: string>
child 0, item: string
properties: struct<record_id: struct<type: string, pattern: string>, task: struct<type: string, required: list<i (... 1150 chars omitted)
child 0, record_id: struct<type: string, pattern: string>
child 0, type: string
child 1, pattern: string
child 1, task: struct<type: string, required: list<item: string>, properties: struct<locale: struct<type: string, m (... 258 chars omitted)
child 0, type: string
child 1, required: list<item: string>
child 0, item: string
child 2, properties: struct<locale: struct<type: string, minLength: int64>, type: struct<type: string, minLength: int64>, (... 166 chars omitted)
child 0, locale: struct<type: string, minLength: int64>
child 0, type: string
child 1, minLength: int64
child 1, type: struct<type: string, minLength: int64>
child 0, type: string
child 1, minLength: int64
child 2, prompt: struct<type: string, minLength: int64>
child 0, type: string
child 1, minLength: int64
child 3, constraints: struct<type: string, minItems: int64, items: struct<type: string, minLength: int64>, uniqueItems: bo (... 3 chars omitted)
child 0, type: string
child 1, minItems: int64
child 2, items: struct<type: string, minLength: int64>
child 0,
...
nstruction_following: int64, n (... 332 chars omitted)
child 0, preferred: string
child 1, strength: string
child 2, scores: struct<A: struct<instruction_following: int64, naturalness: int64, factuality: int64, relevance: int (... 168 chars omitted)
child 0, A: struct<instruction_following: int64, naturalness: int64, factuality: int64, relevance: int64, tone: (... 26 chars omitted)
child 0, instruction_following: int64
child 1, naturalness: int64
child 2, factuality: int64
child 3, relevance: int64
child 4, tone: int64
child 5, conciseness: int64
child 1, B: struct<instruction_following: int64, naturalness: int64, factuality: int64, relevance: int64, tone: (... 26 chars omitted)
child 0, instruction_following: int64
child 1, naturalness: int64
child 2, factuality: int64
child 3, relevance: int64
child 4, tone: int64
child 5, conciseness: int64
child 3, issues: struct<A: list<item: null>, B: list<item: struct<code: string, severity: string>>>
child 0, A: list<item: null>
child 0, item: null
child 1, B: list<item: struct<code: string, severity: string>>
child 0, item: struct<code: string, severity: string>
child 0, code: string
child 1, severity: string
child 4, rationale: string
responses: struct<A: string, B: string>
child 0, A: string
child 1, B: string
record_id: string
to
{'record_id': Value('string'), 'task': {'locale': Value('string'), 'type': Value('string'), 'prompt': Value('string'), 'constraints': List(Value('string'))}, 'responses': {'A': Value('string'), 'B': Value('string')}, 'evaluation': {'preferred': Value('string'), 'strength': Value('string'), 'scores': {'A': {'instruction_following': Value('int64'), 'naturalness': Value('int64'), 'factuality': Value('int64'), 'relevance': Value('int64'), 'tone': Value('int64'), 'conciseness': Value('int64')}, 'B': {'instruction_following': Value('int64'), 'naturalness': Value('int64'), 'factuality': Value('int64'), 'relevance': Value('int64'), 'tone': Value('int64'), 'conciseness': Value('int64')}}, 'issues': {'A': List(Value('null')), 'B': List({'code': Value('string'), 'severity': Value('string')})}, 'rationale': Value('string')}, 'revision_guidance': {'target': Value('string'), 'operations': List(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/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
$schema: string
title: string
type: string
required: list<item: string>
child 0, item: string
properties: struct<record_id: struct<type: string, pattern: string>, task: struct<type: string, required: list<i (... 1150 chars omitted)
child 0, record_id: struct<type: string, pattern: string>
child 0, type: string
child 1, pattern: string
child 1, task: struct<type: string, required: list<item: string>, properties: struct<locale: struct<type: string, m (... 258 chars omitted)
child 0, type: string
child 1, required: list<item: string>
child 0, item: string
child 2, properties: struct<locale: struct<type: string, minLength: int64>, type: struct<type: string, minLength: int64>, (... 166 chars omitted)
child 0, locale: struct<type: string, minLength: int64>
child 0, type: string
child 1, minLength: int64
child 1, type: struct<type: string, minLength: int64>
child 0, type: string
child 1, minLength: int64
child 2, prompt: struct<type: string, minLength: int64>
child 0, type: string
child 1, minLength: int64
child 3, constraints: struct<type: string, minItems: int64, items: struct<type: string, minLength: int64>, uniqueItems: bo (... 3 chars omitted)
child 0, type: string
child 1, minItems: int64
child 2, items: struct<type: string, minLength: int64>
child 0,
...
nstruction_following: int64, n (... 332 chars omitted)
child 0, preferred: string
child 1, strength: string
child 2, scores: struct<A: struct<instruction_following: int64, naturalness: int64, factuality: int64, relevance: int (... 168 chars omitted)
child 0, A: struct<instruction_following: int64, naturalness: int64, factuality: int64, relevance: int64, tone: (... 26 chars omitted)
child 0, instruction_following: int64
child 1, naturalness: int64
child 2, factuality: int64
child 3, relevance: int64
child 4, tone: int64
child 5, conciseness: int64
child 1, B: struct<instruction_following: int64, naturalness: int64, factuality: int64, relevance: int64, tone: (... 26 chars omitted)
child 0, instruction_following: int64
child 1, naturalness: int64
child 2, factuality: int64
child 3, relevance: int64
child 4, tone: int64
child 5, conciseness: int64
child 3, issues: struct<A: list<item: null>, B: list<item: struct<code: string, severity: string>>>
child 0, A: list<item: null>
child 0, item: null
child 1, B: list<item: struct<code: string, severity: string>>
child 0, item: struct<code: string, severity: string>
child 0, code: string
child 1, severity: string
child 4, rationale: string
responses: struct<A: string, B: string>
child 0, A: string
child 1, B: string
record_id: string
to
{'record_id': Value('string'), 'task': {'locale': Value('string'), 'type': Value('string'), 'prompt': Value('string'), 'constraints': List(Value('string'))}, 'responses': {'A': Value('string'), 'B': Value('string')}, 'evaluation': {'preferred': Value('string'), 'strength': Value('string'), 'scores': {'A': {'instruction_following': Value('int64'), 'naturalness': Value('int64'), 'factuality': Value('int64'), 'relevance': Value('int64'), 'tone': Value('int64'), 'conciseness': Value('int64')}, 'B': {'instruction_following': Value('int64'), 'naturalness': Value('int64'), 'factuality': Value('int64'), 'relevance': Value('int64'), 'tone': Value('int64'), 'conciseness': Value('int64')}}, 'issues': {'A': List(Value('null')), 'B': List({'code': Value('string'), 'severity': Value('string')})}, 'rationale': Value('string')}, 'revision_guidance': {'target': Value('string'), 'operations': List(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.
AI Response Evaluation & Revision — Public Sample
This repository contains three seller-authored synthetic evaluation cases for testing response quality, instruction following, naturalness, relevance, tone, conciseness, issue severity, preference decisions, and revision guidance. It is a discovery sample only; the complete paid edition is not included.
Intended uses
- Prototyping LLM evaluator, ranking, and response-revision workflows.
- Demonstrating a structured preference-case schema.
- Qualitative evaluation experiments with appropriate human review.
Unsupported uses
- Treating the records as observed human preferences, production-user behaviour, or benchmark ground truth.
- Using scores as calibrated measurements across unrelated tasks or models.
- Making automated high-stakes decisions without qualified human review.
Contents
sample.jsonl: exactly three records (ARE_P001,ARE_P003,ARE_P006).schema.json: public field contract.PROVENANCE.md: origin and privacy statement.LICENSE: Creative Commons Attribution 4.0 International notice.
The source prototype contains 10 records. A planned Professional edition may contain more records, but no unreleased records are included or promised here.
- Public sample version:
0.1.0 - Source prototype version:
0.1.0 - Release date:
2026-09-01
Provenance and limitations
The records are seller-authored synthetic prompt/response pairs. They are not captured conversations, customer logs, human preference studies, or production annotations. The three examples deliberately cover constrained rewriting, technical troubleshooting, and structured extraction. Candidate ordering is not randomized in this prototype, the sample is too small for statistical conclusions, and multilingual/native-speaker validation is outside this sample.
Full edition
The current commercial product is available on Vertical Marketplace as “AI Response Evaluation & Naturalness Judgment Patterns — Quality, Tone & Instruction Following.”
Full commercial listing: https://verticalmarketplace.ai/data/listing/68f9d8ac-8fdc-4190-8e3b-4227e104cbf9?utm_source=huggingface&utm_medium=dataset_card&utm_campaign=ai_response_eval_sample_v0_1
The link opens the exact verified Vertical Marketplace product page. The paid listing contains the complete currently available product rather than this three-record public discovery sample.
Release status
Published on Hugging Face on 2026-09-01 after content-owner approval and independent QA. The public repository contains only the three-record sample and is licensed under CC BY 4.0.
- Downloads last month
- 33