The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
status: string
all8_models: bool
no_training: bool
no_forwards: bool
independent_corpora: int64
exact_training_objective: string
includes_markers: bool
source_manifest_sha256: string
prior_primary_analysis_unchanged: bool
cpu_threads: int64
soft_stop_seconds: int64
REPORT.md: struct<bytes: int64, sha256: string, git_blob_sha1: string>
child 0, bytes: int64
child 1, sha256: string
child 2, git_blob_sha1: string
summary.json: struct<bytes: int64, sha256: string, git_blob_sha1: string>
child 0, bytes: int64
child 1, sha256: string
child 2, git_blob_sha1: string
protocol.json: struct<bytes: int64, sha256: string, git_blob_sha1: string>
child 0, bytes: int64
child 1, sha256: string
child 2, git_blob_sha1: string
source.py: struct<bytes: int64, sha256: string, git_blob_sha1: string>
child 0, bytes: int64
child 1, sha256: string
child 2, git_blob_sha1: string
to
{'REPORT.md': {'bytes': Value('int64'), 'sha256': Value('string'), 'git_blob_sha1': Value('string')}, 'protocol.json': {'bytes': Value('int64'), 'sha256': Value('string'), 'git_blob_sha1': Value('string')}, 'source.py': {'bytes': Value('int64'), 'sha256': Value('string'), 'git_blob_sha1': Value('string')}, 'summary.json': {'bytes': Value('int64'), 'sha256': Value('string'), 'git_blob_sha1': 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 2951, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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
status: string
all8_models: bool
no_training: bool
no_forwards: bool
independent_corpora: int64
exact_training_objective: string
includes_markers: bool
source_manifest_sha256: string
prior_primary_analysis_unchanged: bool
cpu_threads: int64
soft_stop_seconds: int64
REPORT.md: struct<bytes: int64, sha256: string, git_blob_sha1: string>
child 0, bytes: int64
child 1, sha256: string
child 2, git_blob_sha1: string
summary.json: struct<bytes: int64, sha256: string, git_blob_sha1: string>
child 0, bytes: int64
child 1, sha256: string
child 2, git_blob_sha1: string
protocol.json: struct<bytes: int64, sha256: string, git_blob_sha1: string>
child 0, bytes: int64
child 1, sha256: string
child 2, git_blob_sha1: string
source.py: struct<bytes: int64, sha256: string, git_blob_sha1: string>
child 0, bytes: int64
child 1, sha256: string
child 2, git_blob_sha1: string
to
{'REPORT.md': {'bytes': Value('int64'), 'sha256': Value('string'), 'git_blob_sha1': Value('string')}, 'protocol.json': {'bytes': Value('int64'), 'sha256': Value('string'), 'git_blob_sha1': Value('string')}, 'source.py': {'bytes': Value('int64'), 'sha256': Value('string'), 'git_blob_sha1': Value('string')}, 'summary.json': {'bytes': Value('int64'), 'sha256': Value('string'), 'git_blob_sha1': 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.
Synthetic topic models with a shared persona parent
This dataset contains synthetic checkpoints and numerical evidence from the 10-hour persona/topic-model research sprint of 8–9 October 2026. Public archival was explicitly authorized by the owner after the private storage quota was exhausted. No natural-language training corpus or credentials are included.
Read the research report and construction/EM proposal. The reproduction guide maps claims to code and numerical inputs. A printable report with the six main figures is included with its typesetting source and input hashes.
The controlled models support conditional parent-recovery and adaptation predictions. Their laws depend on observation design, domain geometry and the update rule. A restricted persona-weight fit supports an A/N+B/M reference with learned emissions frozen and domain identities supplied. The neural experiments show lower post-FT prediction error at larger pretraining sizes under the tested schedules, while raw FT benefit and benefit over balanced FT behave differently. These results do not establish a universal transformer scaling law in data size, diversity and one MI ratio.
Artifact layout
- Numerical runs and checkpoints preserve their experiment-relative paths. Failed runs, pilots, runtime gates and independent audits remain identified in the reports.
- Sprint source catalog contains the sprint's own current code, declared historical pipeline versions, source hashes and environment metadata.
- The final report bundle contains the linked reports and figures. Per-run
archive_manifest.jsonfiles record file sizes and SHA256 hashes. Verification records distinguish local SHA256, remote LFS SHA256 and ordinary Git-blob hashes.
Original author source and manuscript files are excluded. The pinned upstream repository returned HTTP404 without authentication during review; authorized access is needed to rerun the author-code baseline or persona transformer training. The standalone structured CPU experiments use the sprint's own archived code. Twenty-seven early neural configs lack the later pipeline-hash field, a documented provenance limitation. Current source is not retroactively assigned to those configurations.
These are exploratory research artifacts with categorical synthetic tokens. Losses use natural logs. Check the report's observation assumptions, uncertainty units and completion record before interpreting a checkpoint or numerical curve.
- Downloads last month
- 50,721