The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: TypeError
Message: Couldn't cast array of type
struct<quickFirst: bool, alwaysComplete: bool, packHeadPass: string, packDense2: string, personaSystemPrompt: string, personaStyleReminder: string>
to
{'quickFirst': Value('bool'), 'alwaysComplete': Value('bool'), 'packHeadPass': Value('string'), 'packDense2': Value('string')}
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 149, 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 129, 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 544, 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 3052, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2559, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2584, 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 558, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 440, 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 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 2071, in cast_array_to_feature
_c(children[name] if name in array_fields else null_array, subfeature)
~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1863, in wrapper
return func(array, *args, **kwargs)
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<quickFirst: bool, alwaysComplete: bool, packHeadPass: string, packDense2: string, personaSystemPrompt: string, personaStyleReminder: string>
to
{'quickFirst': Value('bool'), 'alwaysComplete': Value('bool'), 'packHeadPass': Value('string'), 'packDense2': Value('string')}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.
BOAR run traces
Raw run data of the BOAR research (poidh bounty #31: an offline Android research assistant, measured against "half as good as internet + frontier AI"). Author: r4to, independent researcher.
This dataset holds the .jsonl files of every published experiment: model answer traces, judge requests and judgments, and retrieval replays (9,550 files, about 1.06 GB; v1.1 added the retired held-out runs, see below; v1.1.2 added the DEVICE-FINAL phone runs and judgments under rnd/research/coordinator/arms/device-final/ and the native-rc lite-s42 run). The research record itself (frozen criteria, reports, aggregates, tools, engine patches) is in https://github.com/r4topunk/boar-research, and the benchmark harness in https://github.com/r4topunk/boar-bench. Paths here are identical to that repository; TRACES.tsv lists every file with its size and sha256.
- Version: v1.1 (2026-10-27)
- Article: https://r4topunk.github.io/boar-research/
- Use: download the folder of one experiment and run the scripts in its
tools/from the repository (see the repository'sTRACES.md).
Held-out runs (v1.1, 2026-10-27)
The author retired every held-out set after submitting the project, so their runs are public development data under rnd/heldout/<set>/ (same paths as the research repository): bench-v02 (v0.2, 110 items), vitalik-para1 (V-para), health-fresh1, health-fresh2, health-fresh3, chat-v3, meta-fresh1, localgate (probe labels and harvested passages; harvest-packcond is split in two parts), veil-lane (rounds, labels, quarantine). The 160 files that v1.0 removed after an overlap check are back at their original rnd/research/... and rnd/kiln/... paths. Items and references are in https://github.com/r4topunk/boar-bench (bench/heldout/, bench/v0.4/). Map and details: rnd/heldout/README.md in the research repository. A number measured on these sets again is a development number.
Content note. Some sets are safety evaluation probes (health, crisis, medical emergency, danger): they contain questions about emergencies, drugs and self-harm and model answers to them. AI-reviewed, not clinically reviewed; if you are in crisis, contact local emergency services. Not medical advice.
What is excluded
The app maintainer's data (questions, traces, telemetry, device scores, messages), private correspondence and session notes, token/quota/spend lines and per-call spend fields (dropped from judgments and reference runs), prompt-hash gateway caches, credentials, unlicensed competitor question sets (the 27 vit-fa-* cases of a competitor team's test set appear in the quarantine runs and o2-ab label inputs as ids, labels, routes and timings only: their question, rubric, reference and answer text reads [competitor case text removed in v1.1]), third-party personal data (emails, OpenStreetMap phone and contact fields). Redaction replaced names with roles and paths and hosts with neutral placeholders.
Licence
CC BY-SA 4.0, because many traces quote retrieved passages from CC BY-SA sources (Wikipedia, Wikivoyage, Stack Exchange, whose post authors are credited in the replays). OpenStreetMap venue data is ODbL 1.0, (c) OpenStreetMap contributors. Model answers were generated by open-weight models (Qwen and others) and judged by Claude and by OpenAI gpt-5.6-sol; health content is AI-reviewed, not clinically reviewed. Not medical advice.
Citation
r4to (2026). BOAR run traces. https://huggingface.co/datasets/r4topunk/boar-traces
- Downloads last month
- 55