The dataset viewer is not available for this subset.
Exception: SplitsNotFoundError
Message: The split names could not be parsed from the dataset config.
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 290, in _generate_tables
pa_table = paj.read_json(
io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
)
File "pyarrow/_json.pyx", line 364, in pyarrow._json.read_json
File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
return check_status(status)
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
raise convert_status(status)
pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to string in row 0
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
for split_generator in builder._split_generators(
~~~~~~~~~~~~~~~~~~~~~~~~~^
StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 101, in _split_generators
pa_table = next(iter(self._generate_tables(**splits[0].gen_kwargs, allow_full_read=False)))[1]
~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 304, in _generate_tables
batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
examples = [ujson_loads(line) for line in original_batch.splitlines()]
~~~~~~~~~~~^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
return pd.io.json.ujson_loads(*args, **kwargs)
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
ValueError: Expected object or value
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
for split in get_dataset_split_names(
~~~~~~~~~~~~~~~~~~~~~~~^
path=dataset,
^^^^^^^^^^^^^
config_name=config,
^^^^^^^^^^^^^^^^^^^
token=hf_token,
^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
info = get_dataset_config_info(
path,
...<6 lines>...
**config_kwargs,
)
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.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.
Memory DCI
Minimal reproducibility data for Memory_DCI: V1 and V1.1 baselines, and the V5 task-local memory workflow.
inputs.tar.gz contains the fixed 36-training/24-test split, 432 normalized
training trajectories from 12 models, 381 referenced images, task metadata,
training-score provenance, the 24 test workspaces including grader-only ground
truth, official task-interface skills, and the pinned ripgrep executable with
its upstream license files. Test trajectories never enter the
training pool; the harness mounts ground truth only for grading.
evidence.tar.gz contains the 72 original run results, solver transcripts and
tool events, V5 per-task memory snapshots/reviewer decisions, and historical
protocols. It omits credentials, raw Codex homes/sessions, duplicated output
workspaces, caches, model weights, and retired V2–V4 / Terminal Bench runs.
This is a lean evidence export: full raw-session and workspace integrity audits
cannot be reproduced solely from it. Existing score summaries remain original.
Every archived file has a size and SHA-256 in manifest.json. Credential-like
material is removed before publication; redactions.json records affected
paths, original/export hashes, and replacement categories, never removed values.
The exported corpus manifest records updated hashes and original hashes for
changed trajectories. The sanitized export is not byte-identical to the
original experiment inputs; freshly built pool hashes therefore differ.
Sources
The upstream datasets and benchmark code declare MIT; see provenance.json
and UPSTREAM_LICENSE.txt. Third-party task assets retain their original rights.
This dataset does not assert a new blanket license over those assets.
Restore
From a clone of Memory_DCI, install its requirements and run:
python3 scripts/download_data.py
python3 scripts/download_data.py --include-evidence
python3 scripts/report_results.py
The downloader verifies archive and per-file hashes, restores original relative
paths, stages public test inputs separately from grader-only ground truth, and
can fetch the pinned upstream benchmark using --fetch-benchmark. Docker,
Codex authentication and the runtime image are separate prerequisites documented
in the code repository. No token is required for this public dataset.
- Downloads last month
- 269