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/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/webdataset/webdataset.py", line 84, in _split_generators
raise ValueError(
...<2 lines>...
)
ValueError: The TAR archives of the dataset should be in WebDataset format, but the files in the archive don't share the same prefix or the same types.
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.
Mafia-MD embeddings
Embedding release for 93,882 eligible public-speech clips from 2,937 online Mafia games. Private briefings are excluded. This includes ordinary speeches and eligible re-speeches; task-specific cutoff filters must still be applied for forecasting.
Contents
raw/shards/: original ImageBind audio/video vectors and per-frame GigaAM-Emo / EmotiEff arrays, in uncompressed tar shards. Array bytes are preserved. Tar paths contain modality, game ID and clip stem.compact/: clip-level arrays aligned to therowfield inmetadata/clips.jsonl. ImageBind arrays are float32 vectors; emotion arrays concatenate frame mean and population standard deviation in float32. Text arrays are float16.metadata/clips.jsonl: row, clip ID, game ID, stem, seat, day and speech kind. No transcripts, participant account identifiers or original local paths.metadata/release.json: complete shard list, sizes, SHA256 checksums and exact feature recipes.
Text coverage
e5_noprefix.npy contains existing multilingual-e5-large pooled text features without a prefix, L2 normalized, where cached, with benchmark no-prefix values reused as a fallback. Apply e5_noprefix_available.npy to exclude unavailable rows. e5.npy reuses the benchmark's separately calculated query-prefix features where available. Query-prefix coverage is partial: rows outside that benchmark subset are zero and must be excluded using e5_query_available.npy. A zero row is not a valid embedding.
Loading compact features
import numpy as np
from huggingface_hub import hf_hub_download
path = hf_hub_download("SODA2207/MafiaMD", "compact/ib_audio.npy", repo_type="dataset")
features = np.load(path, mmap_mode="r", allow_pickle=False)
Raw emotion arrays retain frame-level information; compact arrays pool it. This release does not include the later 400-game replication cohort or all private inputs needed for every experiment. Do not attempt to identify participants or link their representations to real identities. Access remains subject to the dataset's access conditions; no additional media redistribution rights are granted by this feature release.
Upload complete. Validate downloads against the SHA256 values in the release manifest.
- Downloads last month
- 236