The metadata for a dataset includes:

• language (en, ja, etc)
• task (analogy, similarity, etc)
• description (e.g. Bigger Analogy Test Set)
• version (e.g. 3.0)
• cite (bibtex for the paper to cite)

Word similarity:

1. WordSim 353
2. MEN
3. SimLex
4. Rare Words
5. MTurk

Word analogy:

1. BATS

Text classification:

1. IMDb moview reviews sample

Word similarity:

1. Japanese word similarity (https://github.com/tmu-nlp/JapaneseWordSimilarityDataset)

Japanese word similarity:

1. JBATS