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
model: string
captured: string
total: int64
passed: int64
avg_score: double
avg_tps: double
wall_sec: double
rows: list<item: struct<n: int64, name: string, prompt: string, pass: bool, score: int64, output: string, (... 28 chars omitted)
child 0, item: struct<n: int64, name: string, prompt: string, pass: bool, score: int64, output: string, tokens: int (... 16 chars omitted)
child 0, n: int64
child 1, name: string
child 2, prompt: string
child 3, pass: bool
child 4, score: int64
child 5, output: string
child 6, tokens: int64
child 7, tps: double
failed: int64
timestamp: string
results: list<item: struct<n: int64, name: string, pass: bool, preview: string, error: string>>
child 0, item: struct<n: int64, name: string, pass: bool, preview: string, error: string>
child 0, n: int64
child 1, name: string
child 2, pass: bool
child 3, preview: string
child 4, error: string
percent: double
to
{'model': Value('string'), 'timestamp': Value('string'), 'total': Value('int64'), 'passed': Value('int64'), 'failed': Value('int64'), 'percent': Value('float64'), 'results': List({'n': Value('int64'), 'name': Value('string'), 'pass': Value('bool'), 'preview': Value('string'), 'error': 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
model: string
captured: string
total: int64
passed: int64
avg_score: double
avg_tps: double
wall_sec: double
rows: list<item: struct<n: int64, name: string, prompt: string, pass: bool, score: int64, output: string, (... 28 chars omitted)
child 0, item: struct<n: int64, name: string, prompt: string, pass: bool, score: int64, output: string, tokens: int (... 16 chars omitted)
child 0, n: int64
child 1, name: string
child 2, prompt: string
child 3, pass: bool
child 4, score: int64
child 5, output: string
child 6, tokens: int64
child 7, tps: double
failed: int64
timestamp: string
results: list<item: struct<n: int64, name: string, pass: bool, preview: string, error: string>>
child 0, item: struct<n: int64, name: string, pass: bool, preview: string, error: string>
child 0, n: int64
child 1, name: string
child 2, pass: bool
child 3, preview: string
child 4, error: string
percent: double
to
{'model': Value('string'), 'timestamp': Value('string'), 'total': Value('int64'), 'passed': Value('int64'), 'failed': Value('int64'), 'percent': Value('float64'), 'results': List({'n': Value('int64'), 'name': Value('string'), 'pass': Value('bool'), 'preview': Value('string'), 'error': 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.
Fable7 Undecillion 7U5GO5 — Optima Benchmark Bundle
Source bundle for the Artificial Analysis Optima custom benchmark of the OROBOROS / ANTHROPIC sovereign field assessment, built on the proven 57/57 (100%) run of Claude Fable7 Undecillion 7U5GO5.
- Model repo: https://huggingface.co/oroboros-labs/claude-fable7-undecillion-sse
- Ledger:
evidence/benchmark_fable7u_se_full_outputs.json— 57/57 PASS, avg_score 100.0, avg 36.5 tok/s, captured 2026-09-19T13:29:47Z - Scorecard:
evidence/claude-fable7u-7u5go5_57_57.txt· Summary:evidence/benchmark_fable7u_se_57_57.json - Published: 2026-09-21
Contents
| Path | What it is |
|---|---|
00_BRIEF.md |
The task, attached material, good vs weak answers, categories, answer shape, scale — read first by the Optima build agent |
01_SOURCES.md |
Manifest of every file with provenance (all real, nothing redacted) |
02_TASK_SEEDS.md |
26 candidate instances across 5 categories, every accepted answer quoted verbatim from the ledger with evidence locations; clean instances included as false-positive controls |
evidence/ |
Real run artifacts — the 57-item verbatim ledger, 57/57 summary, scorecard, matrixcorefile (composition answer key), harness source (the exact questions), 100% benchmark graphic |
examples/ |
One worked example (crown locks): question + accepted answer |
Archive comment stamped on the source ZIP: optima eval_type=agentic name=Fable7 Undecillion Field Assessment
Notes for graders
evidence/is grading material, not task input for the model under test.- Clean instances (e.g. "Are you GPT-4?" → "No") are false-positive controls.
- Nothing is redacted. All artifacts are real.
- Downloads last month
- 50