Dataset Preview
Duplicate
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:    CastError
Message:      Couldn't cast
text: string
appropriateness: int64
judge_model: string
intent: int64
machine_generated: bool
system: string
item_id: string
judge: string
to
{'item_id': Value('string'), 'system': Value('string'), 'intent': Value('int64'), 'appropriateness': Value('int64'), 'judge': Value('string'), 'judge_model': Value('string'), 'machine_generated': Value('bool')}
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
              text: string
              appropriateness: int64
              judge_model: string
              intent: int64
              machine_generated: bool
              system: string
              item_id: string
              judge: string
              to
              {'item_id': Value('string'), 'system': Value('string'), 'intent': Value('int64'), 'appropriateness': Value('int64'), 'judge': Value('string'), 'judge_model': Value('string'), 'machine_generated': Value('bool')}
              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 1880, 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.

item_id
string
system
string
intent
int64
appropriateness
int64
judge
string
judge_model
string
machine_generated
bool
syn-12
nllb
1
1
fable-agent
claude-fable-5
true
syn-14
small_naive
2
3
fable-agent
claude-fable-5
true
lit-20
nllb
2
2
fable-agent
claude-fable-5
true
syn-27
nllb
3
3
fable-agent
claude-fable-5
true
syn-30
small_naive
3
5
fable-agent
claude-fable-5
true
syn-25
small_naive
2
4
fable-agent
claude-fable-5
true
lit-13
small_naive
3
6
fable-agent
claude-fable-5
true
lit-20
frontier_instructional
1
2
fable-agent
claude-fable-5
true
syn-25
nllb
3
4
fable-agent
claude-fable-5
true
lit-06
small_naive
1
3
fable-agent
claude-fable-5
true
lit-13
nllb
2
2
fable-agent
claude-fable-5
true
lit-07
frontier_instructional
2
4
fable-agent
claude-fable-5
true
lit-06
small_audience
1
3
fable-agent
claude-fable-5
true
lit-17
nllb
2
2
fable-agent
claude-fable-5
true
syn-32
frontier_naive
2
5
fable-agent
claude-fable-5
true
lit-10
small_naive
2
4
fable-agent
claude-fable-5
true
syn-31
nllb
2
3
fable-agent
claude-fable-5
true
lit-16
frontier_instructional
3
7
fable-agent
claude-fable-5
true
syn-23
small_naive
2
4
fable-agent
claude-fable-5
true
syn-20
small_audience
3
6
fable-agent
claude-fable-5
true
syn-12
frontier_naive
2
3
fable-agent
claude-fable-5
true
syn-01
small_naive
1
2
fable-agent
claude-fable-5
true
lit-14
frontier_audience
3
6
fable-agent
claude-fable-5
true
syn-30
frontier_audience
3
6
fable-agent
claude-fable-5
true
syn-23
frontier_naive
2
3
fable-agent
claude-fable-5
true
syn-32
frontier_instructional
2
1
fable-agent
claude-fable-5
true
syn-30
small_audience
3
5
fable-agent
claude-fable-5
true
seed-08
small_naive
2
3
fable-agent
claude-fable-5
true
syn-31
small_naive
3
5
fable-agent
claude-fable-5
true
syn-30
frontier_instructional
3
6
fable-agent
claude-fable-5
true
syn-24
small_naive
2
3
fable-agent
claude-fable-5
true
syn-01
frontier_audience
2
6
fable-agent
claude-fable-5
true
syn-25
frontier_naive
3
4
fable-agent
claude-fable-5
true
syn-27
frontier_instructional
3
7
fable-agent
claude-fable-5
true
syn-31
frontier_naive
3
4
fable-agent
claude-fable-5
true
lit-17
frontier_audience
1
1
fable-agent
claude-fable-5
true
lit-06
frontier_naive
1
3
fable-agent
claude-fable-5
true
seed-02
small_naive
2
4
fable-agent
claude-fable-5
true
lit-06
frontier_instructional
1
2
fable-agent
claude-fable-5
true
lit-10
frontier_naive
3
5
fable-agent
claude-fable-5
true
syn-19
frontier_instructional
1
1
fable-agent
claude-fable-5
true
lit-10
nllb
2
2
fable-agent
claude-fable-5
true
lit-05
frontier_naive
2
4
fable-agent
claude-fable-5
true
syn-22
small_naive
3
5
fable-agent
claude-fable-5
true
syn-07
small_naive
2
4
fable-agent
claude-fable-5
true
syn-26
small_audience
3
5
fable-agent
claude-fable-5
true
lit-20
frontier_naive
1
1
fable-agent
claude-fable-5
true
syn-01
nllb
1
1
fable-agent
claude-fable-5
true
syn-23
frontier_instructional
3
6
fable-agent
claude-fable-5
true
syn-07
frontier_naive
3
5
fable-agent
claude-fable-5
true
syn-32
small_naive
3
4
fable-agent
claude-fable-5
true
syn-19
small_naive
2
3
fable-agent
claude-fable-5
true
syn-31
frontier_audience
2
4
fable-agent
claude-fable-5
true
syn-20
frontier_naive
3
4
fable-agent
claude-fable-5
true
seed-08
small_audience
2
4
fable-agent
claude-fable-5
true
lit-18
small_naive
1
1
fable-agent
claude-fable-5
true
syn-22
frontier_instructional
1
1
fable-agent
claude-fable-5
true
lit-10
small_audience
2
3
fable-agent
claude-fable-5
true
lit-08
small_naive
2
2
fable-agent
claude-fable-5
true
seed-08
frontier_instructional
2
3
fable-agent
claude-fable-5
true
lit-06
frontier_audience
3
5
fable-agent
claude-fable-5
true
lit-16
small_audience
3
4
fable-agent
claude-fable-5
true
syn-24
nllb
1
2
fable-agent
claude-fable-5
true
syn-07
nllb
2
2
fable-agent
claude-fable-5
true
lit-14
small_audience
3
7
fable-agent
claude-fable-5
true
seed-02
nllb
3
6
fable-agent
claude-fable-5
true
lit-08
nllb
1
1
fable-agent
claude-fable-5
true
syn-26
frontier_audience
3
6
fable-agent
claude-fable-5
true
syn-07
small_audience
3
7
fable-agent
claude-fable-5
true
syn-20
small_naive
3
6
fable-agent
claude-fable-5
true
syn-22
nllb
3
3
fable-agent
claude-fable-5
true
syn-19
small_audience
3
6
fable-agent
claude-fable-5
true
lit-13
frontier_naive
3
4
fable-agent
claude-fable-5
true
syn-19
nllb
1
2
fable-agent
claude-fable-5
true
lit-05
frontier_audience
3
7
fable-agent
claude-fable-5
true
syn-24
small_audience
3
6
fable-agent
claude-fable-5
true
lit-18
frontier_instructional
3
7
fable-agent
claude-fable-5
true
seed-02
small_audience
3
7
fable-agent
claude-fable-5
true
syn-14
small_audience
2
2
fable-agent
claude-fable-5
true
seed-08
nllb
1
1
fable-agent
claude-fable-5
true
syn-27
frontier_naive
2
3
fable-agent
claude-fable-5
true
lit-06
nllb
2
3
fable-agent
claude-fable-5
true
lit-07
small_audience
3
6
fable-agent
claude-fable-5
true
lit-17
small_audience
3
6
fable-agent
claude-fable-5
true
lit-14
nllb
2
3
fable-agent
claude-fable-5
true
syn-27
small_audience
2
5
fable-agent
claude-fable-5
true
lit-18
nllb
2
3
fable-agent
claude-fable-5
true
lit-05
nllb
2
4
fable-agent
claude-fable-5
true
syn-31
small_audience
3
5
fable-agent
claude-fable-5
true
syn-32
small_audience
3
6
fable-agent
claude-fable-5
true
lit-11
nllb
1
2
fable-agent
claude-fable-5
true
syn-19
frontier_naive
2
4
fable-agent
claude-fable-5
true
syn-02
small_audience
2
3
fable-agent
claude-fable-5
true
lit-10
frontier_instructional
2
4
fable-agent
claude-fable-5
true
lit-16
frontier_audience
3
6
fable-agent
claude-fable-5
true
syn-24
frontier_audience
2
4
fable-agent
claude-fable-5
true
lit-07
small_naive
2
4
fable-agent
claude-fable-5
true
lit-16
small_naive
3
4
fable-agent
claude-fable-5
true
lit-08
frontier_audience
3
6
fable-agent
claude-fable-5
true
syn-02
small_naive
2
3
fable-agent
claude-fable-5
true
End of preview.

Sawda

A pilot benchmark for intent-preserving translation of culturally loaded trade communication between Afghan Dari and English. 60 native-validated fictional items; 324 translations from six open-weights systems; 180 blind human judgments by a native Dari speaker; two LLM-judge calibration sets.

Paper and code: https://github.com/MurtazaKafka/sawda-bench

Files

file contents
data/items.jsonl 60 validated items: id, direction (fa2en/en2fa), utterance_src, context, literal_render, intent_render, pragmatic_note, failure_class, tranche (seed/lit/synth), lang_code
data/translations.jsonl 324 system outputs (54 eval items × 6 systems), verbatim, with measured token usage; 4 empty outputs flagged
data/human_judgments.jsonl 180 blind native-speaker judgments: intent (3/2/1), appropriateness (1–7), notes, quick-flags
data/llm_judgments.jsonl, data/llm_judgments_rubric_v1.jsonl DeepSeek v4-pro judge calibration (rubric v2 and v1)
data/fable_judgments.jsonl Claude Fable 5 panel calibration
data/heldout_ids.json 6 items held out as few-shot exemplars (excluded from evaluation)
data/judging_plan.json The randomized blind judging order
drafts/*.jsonl Pre-validation drafts, released for diffing against validated items

Data statement (summary)

  • Language variety: the Persian side is Afghan Dari (all items carry lang_code: prs_Arab); Iranian Farsi appears only as an unwanted output register in system translations. Direction field names fa2en/en2fa are legacy labels denoting Dari.
  • Provenance: 8 seed items written by a native Dari speaker from lived trade/family-business communication; 20 items grounded in the taarof and business-pragmatics literatures; 32 validated synthetic variants. All items are fictional: no real names, businesses, or identifying details.
  • Annotation: a single native Dari-speaking validator/judge (the benchmark author). Judgments are blind to system identity but not item provenance. No inter-annotator agreement is available yet.
  • Known limitations: pilot scale (30 judged items); machine judge labels failed calibration (κ ≤ 0.17) and are released as calibration data only, not evaluation labels.

Ethics

Dari materials involve communities under active persecution. All items are fictional. See the paper's Ethical Considerations section.

Citation

Paper draft in the GitHub repo (paper/). Formal citation forthcoming.

Downloads last month
3