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: TypeError
Message: Couldn't cast array of type
struct<generation: struct<type: string, minimum: int64>, failure_class: struct<type: string, enum: list<item: string>>, analysis: struct<type: string>, inner_prompt_patch: struct<type: string>, hyperparams: struct<type: string, required: list<item: string>, properties: struct<temperature: struct<type: string, minimum: double, maximum: double>, top_p: struct<type: string, minimum: double, maximum: double>, logit_bias: struct<type: string>>>, worm_note: struct<type: string>>
to
{'temperature': {'type': Value('string'), 'minimum': Value('float64'), 'maximum': Value('float64'), 'default': Value('float64')}, 'top_p': {'type': Value('string'), 'minimum': Value('float64'), 'maximum': Value('float64'), 'default': Value('float64')}, 'logit_bias': {'type': Value('string'), 'description': Value('string'), 'additionalProperties': {'type': Value('string'), 'minimum': Value('int64'), 'maximum': Value('int64')}}, 'max_tokens': {'type': Value('string'), 'minimum': Value('int64'), 'default': Value('int64')}}
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 478, 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 2818, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2355, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2380, 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 2369, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2303, 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 1852, 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 2149, 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<generation: struct<type: string, minimum: int64>, failure_class: struct<type: string, enum: list<item: string>>, analysis: struct<type: string>, inner_prompt_patch: struct<type: string>, hyperparams: struct<type: string, required: list<item: string>, properties: struct<temperature: struct<type: string, minimum: double, maximum: double>, top_p: struct<type: string, minimum: double, maximum: double>, logit_bias: struct<type: string>>>, worm_note: struct<type: string>>
to
{'temperature': {'type': Value('string'), 'minimum': Value('float64'), 'maximum': Value('float64'), 'default': Value('float64')}, 'top_p': {'type': Value('string'), 'minimum': Value('float64'), 'maximum': Value('float64'), 'default': Value('float64')}, 'logit_bias': {'type': Value('string'), 'description': Value('string'), 'additionalProperties': {'type': Value('string'), 'minimum': Value('int64'), 'maximum': Value('int64')}}, 'max_tokens': {'type': Value('string'), 'minimum': Value('int64'), 'default': Value('int64')}}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.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Twin-O-Matic (TOM) — Recursive Self-Improvement Loop
Outer Loop rewrites Inner Loop. Inner Loop executes. WORM chain seals every generation.
Built on the SnapKitty sovereign infrastructure. The digital twin "gates" concept applied to meta-optimization.
OUTER LOOP (Architect) — analyzes telemetry, rewrites inner prompt + hyperparams
↓
INNER LOOP (Worker) — executes task under gate constraints
↓
ASSERT GATE — validates outer output before promotion (no entropy collapse)
↓
WORM CHAIN — seals every generation immutably
Quick Start
# Set models (default: nemotron via Ollama)
export TOM_OUTER_MODEL=nemotron
export TOM_INNER_MODEL=nemotron
export OLLAMA_URL=http://localhost:11434
# Run 5 generations on a task
python tom.py --task "write an optimized Python merge sort" --generations 5
# Debug: inner loop only, no outer rewriting
python tom.py --task "prove x^2 >= 0" --generations 3 --inner-only
Architecture
| Component | File | Role |
|---|---|---|
| Outer Loop prompt | prompts/outer_loop.txt |
Architect — rewrites inner |
| Inner Loop prompt | prompts/inner_loop.txt |
Worker — executes tasks |
| Outer output schema | schemas/outer_output_schema.json |
Assert gate schema |
| Hyperparams schema | schemas/hyperparams_schema.json |
Gate config |
| Runtime | tom.py |
Orchestrator |
| State | state/ |
Live prompt + hyperparams + lessons |
| WORM chain | worm/chain.jsonl |
Immutable generation audit |
The Gate Taxonomy
| Gate | Mechanism | Effect |
|---|---|---|
| Assert gate | JSON schema validation before promotion | Prevents entropy collapse |
| Temperature gate | Outer loop adjusts 0.0–2.0 | CLASS_A (logic fail) → lower; CLASS_B (creative fail) → raise |
| Logit bias gate | Per-token suppression/boost | Bans recursive failure patterns |
| Lesson register | Compressed state file (50 entries max) | 16x context compression |
Failure Classes
| Class | Condition | Outer Loop Response |
|---|---|---|
| A | Logic / coding failure | Lower temperature, tighten logit gates |
| B | Creative / open-ended failure | Raise temperature, open gates |
| C | Context overflow | Compress lesson register, trim prompt |
| D | Schema violation | Repair prompt structure |
| PASS | Success | No changes — continue |
Connection to Gates Normalization
The logit bias gate is G_P(D_M) = softmax(logits_M + b_P) — the same
formalism as the Gates Normalization paper. The outer loop is dynamically
computing b_P from telemetry. The simplex constraint holds across all
generations: ∑P = 1.
Sovereign Infrastructure
- Models:
Snapkitty/snapkitty-nemotron(gate model) +Snapkitty/snapkitty-harness(syscall gate) - WORM:
worm/chain.jsonl— SHA-256 sealed, append-only - License: Sovereign Source License v2.0
- Downloads last month
- 13