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
made: string
family: string
from: int64
chars: int64
comma: int64
commaAll: list<item: int64>
child 0, item: int64
dot: int64
sections: list<item: null>
child 0, item: null
marks: list<item: null>
child 0, item: null
model: string
max: list<item: int64>
child 0, item: int64
engine: struct<from: int64, bins: int64, per: double, max: list<item: int64>>
child 0, from: int64
child 1, bins: int64
child 2, per: double
child 3, max: list<item: int64>
child 0, item: int64
bins: int64
per: double
deepest: int64
to
{'made': Value('string'), 'model': Value('string'), 'bins': Value('int64'), 'per': Value('float64'), 'max': List(Value('int64')), 'engine': {'from': Value('int64'), 'bins': Value('int64'), 'per': Value('float64'), 'max': List(Value('int64'))}, 'deepest': Value('int64')}
because column names don't match
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 2309, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
made: string
family: string
from: int64
chars: int64
comma: int64
commaAll: list<item: int64>
child 0, item: int64
dot: int64
sections: list<item: null>
child 0, item: null
marks: list<item: null>
child 0, item: null
model: string
max: list<item: int64>
child 0, item: int64
engine: struct<from: int64, bins: int64, per: double, max: list<item: int64>>
child 0, from: int64
child 1, bins: int64
child 2, per: double
child 3, max: list<item: int64>
child 0, item: int64
bins: int64
per: double
deepest: int64
to
{'made': Value('string'), 'model': Value('string'), 'bins': Value('int64'), 'per': Value('float64'), 'max': List(Value('int64')), 'engine': {'from': Value('int64'), 'bins': Value('int64'), 'per': Value('float64'), 'max': List(Value('int64'))}, 'deepest': Value('int64')}
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.
The brainfuck programs of https://brnfck.chat, and the models trained for them. Every program is one text file of nothing but the eight commands.
Programs that run a chat model
bf/, each with its meta.json:
| Program | Model | Characters |
|---|---|---|
brnfck-smollm2-135m.bf |
SmolLM2-135M-Instruct | 1,030,902,380 |
brnfck-qwen2.5-0.5b.bf |
Qwen2.5-0.5B-Instruct | 4,169,659,608 |
brnfck-qwen2.5-1.5b.bf |
Qwen2.5-1.5B-Instruct | 11,430,468,880 |
brnfck-qwen3.5-0.8b.bf |
Qwen3.5-0.8B | 7,709,034,855 |
brnfck-qwen3.5-0.8b-vision.bf |
Qwen3.5-0.8B, with its part for pictures | 8,931,101,427 |
brnfck-qwen2.5-coder-0.5b.bf |
models/qwen2.5-coder-0.5b-brainfuck |
4,221,543,730 |
programs/<id>/: the same programs as the compressed pieces the Chatbot reads.inside/<id>/: the data of the pages Inside the program and Program map.
Programs that train a network
bf/, each with its meta.json. Each starts from random numbers, learns from what it reads through
its comma command, and then prints a second brainfuck program, its child, with the trained weights in it.
| Program | Trains | Characters |
|---|---|---|
brnfck-train-names-bigram-756.bf |
a table of letter pairs, on names | 1,246,022 |
brnfck-train-names-mlp-7k.bf |
64 hidden units over the last three letters, on names | 2,956,550 |
brnfck-train-names-attn-10k.bf |
the same, plus attention over the last eight letters, on names | 9,023,355 |
brnfck-train-shakespeare-attn-55k.bf |
attention over the last 16 characters, on Shakespeare | 28,992,978 |
bf/children/: what they printed.brnfck-names-bigram-756.bf,brnfck-names-mlp-7k.bfandbrnfck-names-attn-10k.bfmake up names;brnfck-shakespeare-attn-55k.bf, pass 18 of the run below, writes lines like Shakespeare's.bf/generations/: training programs that print their next generation.brnfck-train-names-<net>-gen0.bftrains on the first part of the names and printsgen1: itself, with the weights it ended with written in.gen1trains on the next part and printsgen2, and so on.runs/brnfck-shakespeare-attn-55k/: the whole Shakespeare run, 19 passes, 596,030 steps, 13 h 53 min in the WebAssembly runtime. The child of every pass (pass-NN.bf) and what it wrote (pass-NN.sample.txt), the log of the run (log.txt), the state it ended with (state.json), the weights of the last two passes (pass-18.json,pass-19.json, and their tapes) and the summary (summary.txt).
Models trained for brnfck-qwen2.5-coder-0.5b
models/<name>/: Qwen2.5-Coder-0.5B-Instruct trained further to answer with brainfuck, in the layout
of the original download (model.safetensors in bf16, the config and the tokenizer), each with its
LICENSE and a NOTICE that says how it was changed:
qwen2.5-coder-0.5b-stack: after reading the 2,350 brainfuck programs of The Stack (bigcode/the-stack-dedup; the programs themselves are not in this dataset).qwen2.5-coder-0.5b-brainfuck-v1: the first round, 60,000 examples of 35 kinds of tasks.qwen2.5-coder-0.5b-brainfuck: the final round, The Stack first and then 160,000 examples of 49 kinds; the model ofbrnfck-qwen2.5-coder-0.5b.bf. It writes a passing program for 30 of 60 blind tasks and 32 of 51 challenges.
Code: https://github.com/rutgerklamer/brnfck-chat.
The programs hold the weights of SmolLM2-135M-Instruct (Hugging Face), Qwen2.5-0.5B-Instruct,
Qwen2.5-1.5B-Instruct, Qwen2.5-Coder-0.5B-Instruct and Qwen3.5-0.8B (Alibaba Cloud), and the models in
models/ are Qwen2.5-Coder-0.5B-Instruct trained further; all Apache 2.0: see bf/NOTICE.txt and
bf/LICENSE-Apache-2.0.txt, and models/<name>/NOTICE and LICENSE.
- Downloads last month
- 318