Datasets:

Modalities:
Text
Formats:
csv
Languages:
English
ArXiv:
License:
Dataset Viewer
Auto-converted to Parquet Duplicate
variable_name
int64
timestamps
string
values
string
variable_type
string
0
["1970-01-01T00:00:00Z", "1970-01-02T00:00:00Z", "1970-01-03T00:00:00Z", "1970-01-04T00:00:00Z", "1970-01-05T00:00:00Z", "1970-01-06T00:00:00Z", "1970-01-07T00:00:00Z", "1970-01-08T00:00:00Z", "1970-01-09T00:00:00Z", "1970-01-10T00:00:00Z", "1970-01-11T00:00:00Z", "1970-01-12T00:00:00Z", "1970-01-13T00:00:00Z", "1970-0...
[0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
target

TempusBench — task data

Time-series task CSVs in TempusBench format (variable_name, timestamps, values, variable_type). Each config is one benchmark task folder (see YAML configs in the frontmatter above).

Citation

TempusBench: https://arxiv.org/pdf/2604.11529

License

This dataset card applies to the bundled files as released on the Hub. Project default license for TempusBench: CC-BY-NC-ND-4.0. Individual series may inherit upstream terms — see paper and source listings.


Provenance & Responsible AI (NeurIPS / Croissant placeholders)

⚠️ PLACEHOLDER — fill before Croissant / OpenReview: prov:wasDerivedFrom (upstream URIs per series), rai:dataLimitations, rai:dataBiases, rai:personalSensitiveInformation.

If these are not yet finalized, export Core Croissant from the Hugging Face API after upload, then complete fields in the Croissant RAI editor.

Where to pull source URIs: arXiv text and supplementary materials for TempusBench first; then manual curation for any gaps.


Downloads last month
18

Paper for Smlcrm/tempus_bench