The dataset viewer is not available for this split.
Error code: TooBigContentError
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.
REPLICA study bundles
Versioned inputs for released REPLICA studies. Each archive has a complete
SHA-256 and byte-length inventory in MANIFEST.json; library indexes pin the
archive digest and Hub revision.
The Hopton unigram bundle contains retained predictor estimates, participant reading times, fold assignments and runtime measurements. The archive preserves the original file bytes. Scientific validation is documented in the corresponding study.
The reading-time inputs derive from the Multilingual Eye-movement Corpus (MECO), L1 release 1.0 / version 1.2. The source dataset is provided under CC BY 4.0. The bundled participant table is preprocessed and includes trial-identity corrections for the study analyses. See the corresponding study documentation for preparation and estimator provenance.
Study instructions and provenance are in the Hopton unigram documentation.
| Archive SHA-256 prefix | Contents |
|---|---|
3f59ddf69045 |
Analysis inputs and folds |
321a00826389 |
Variance inputs and reference code |
6efefa6769a9 |
Manuscript table excerpts |
The speech bundles contain processed pupil measurements, predictor values and probability inputs. Raw audio is not included.
| Archive SHA-256 prefix | Contents |
|---|---|
ea5dd3a75cda |
Speech measurements and predictors |
2ce0b47d2946 |
Speech probability inputs |
- Downloads last month
- 130