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
hidden: struct<last_tok: struct<open: struct<mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: d (... 3004 chars omitted)
child 0, last_tok: struct<open: struct<mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64 (... 508 chars omitted)
child 0, open: struct<mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: str (... 192 chars omitted)
child 0, mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: string>
child 0, anchor_A: double
child 1, n_a: int64
child 2, consensus_A: double
child 3, n_c: int64
child 4, strong: string
child 1, cka: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: string>
child 0, anchor_A: double
child 1, n_a: int64
child 2, consensus_A: double
child 3, n_c: int64
child 4, strong: string
child 2, cosine: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: string>
child 0, anchor_A: double
child 1, n_a: int64
child 2, consensus_A: double
child 3, n_c: int64
child 4, strong: string
child 1, closed: struct<mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: str (... 192 chars omitted)
child 0,
...
st<item: st (... 247 chars omitted)
child 0, n_problems: int64
child 1, panels: struct<open: struct<axis: string, rho: double, topN: list<item: string>, points: list<item: struct<m (... 212 chars omitted)
child 0, open: struct<axis: string, rho: double, topN: list<item: string>, points: list<item: struct<model: string, (... 44 chars omitted)
child 0, axis: string
child 1, rho: double
child 2, topN: list<item: string>
child 0, item: string
child 3, points: list<item: struct<model: string, cap: double, align: double, is_top: bool>>
child 0, item: struct<model: string, cap: double, align: double, is_top: bool>
child 0, model: string
child 1, cap: double
child 2, align: double
child 3, is_top: bool
child 1, closed: struct<axis: string, rho: double, topN: list<item: string>, points: list<item: struct<model: string, (... 44 chars omitted)
child 0, axis: string
child 1, rho: double
child 2, topN: list<item: string>
child 0, item: string
child 3, points: list<item: struct<model: string, cap: double, align: double, is_top: bool>>
child 0, item: struct<model: string, cap: double, align: double, is_top: bool>
child 0, model: string
child 1, cap: double
child 2, align: double
child 3, is_top: bool
to
{'last_tok': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}, 'mean_last': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}, 'layer32': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}, 'layer48': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}, 'layer64': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}}
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
hidden: struct<last_tok: struct<open: struct<mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: d (... 3004 chars omitted)
child 0, last_tok: struct<open: struct<mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64 (... 508 chars omitted)
child 0, open: struct<mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: str (... 192 chars omitted)
child 0, mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: string>
child 0, anchor_A: double
child 1, n_a: int64
child 2, consensus_A: double
child 3, n_c: int64
child 4, strong: string
child 1, cka: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: string>
child 0, anchor_A: double
child 1, n_a: int64
child 2, consensus_A: double
child 3, n_c: int64
child 4, strong: string
child 2, cosine: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: string>
child 0, anchor_A: double
child 1, n_a: int64
child 2, consensus_A: double
child 3, n_c: int64
child 4, strong: string
child 1, closed: struct<mutual_knn: struct<anchor_A: double, n_a: int64, consensus_A: double, n_c: int64, strong: str (... 192 chars omitted)
child 0,
...
st<item: st (... 247 chars omitted)
child 0, n_problems: int64
child 1, panels: struct<open: struct<axis: string, rho: double, topN: list<item: string>, points: list<item: struct<m (... 212 chars omitted)
child 0, open: struct<axis: string, rho: double, topN: list<item: string>, points: list<item: struct<model: string, (... 44 chars omitted)
child 0, axis: string
child 1, rho: double
child 2, topN: list<item: string>
child 0, item: string
child 3, points: list<item: struct<model: string, cap: double, align: double, is_top: bool>>
child 0, item: struct<model: string, cap: double, align: double, is_top: bool>
child 0, model: string
child 1, cap: double
child 2, align: double
child 3, is_top: bool
child 1, closed: struct<axis: string, rho: double, topN: list<item: string>, points: list<item: struct<model: string, (... 44 chars omitted)
child 0, axis: string
child 1, rho: double
child 2, topN: list<item: string>
child 0, item: string
child 3, points: list<item: struct<model: string, cap: double, align: double, is_top: bool>>
child 0, item: struct<model: string, cap: double, align: double, is_top: bool>
child 0, model: string
child 1, cap: double
child 2, align: double
child 3, is_top: bool
to
{'last_tok': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}, 'mean_last': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}, 'layer32': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}, 'layer48': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}, 'layer64': {'n_problems': Value('int64'), 'panels': {'open': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}, 'closed': {'axis': Value('string'), 'rho': Value('float64'), 'topN': List(Value('string')), 'points': List({'model': Value('string'), 'cap': Value('float64'), 'align': Value('float64'), 'is_top': Value('bool')})}}}}
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.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Refanchor Platonic Hidden-state Backup
Private archive of long-running Qwen3 anchor-encoded features for Platonic-hypothesis analysis.
Layout
hidden3_235b/<model>.jsonl # Qwen3-235B-A22B encoded, 4096-d, 7 layers, top-32 logit-dist
hidden3_32b/<model>.jsonl # Qwen3-32B encoded, 5120-d, 7 layers, top-32 logit-dist
gen/<model>.jsonl # OPEN-panel model answers (OLLB models)
gen_closed/<model>.jsonl # CLOSED-panel model answers (AA Index models)
gold/refanchor_gold.jsonl # gold problems
analysis/*.json # computed Platonic metrics (small)
Per-row schema (hidden3 jsonl)
{
"id": "...", "domain": "...",
"last_tok": [..H..], # final layer last response token
"mean_last": [..H..], # final layer mean-pool over response
"layer_means": {"L":[..H..]}, # mid layers (32B: 16-64; 235B: 24-94)
"logit_dist": {"<vocab_id>": prob} # top-32 per token, mean over response
}
Anchors: 32B / 235B forward over (prompt+response). 34 scored models = 20 OPEN (OLLB) + 14 CLOSED (AA Index).
- Downloads last month
- 418