The columns are sortable by clicking on the sortable picture of each column header. A detailed view of the results is available by clicking on the details picture of each row.

The columns are interpreted as follows (see Evaluation metrics for details):

  • Phonetic (across and within)

    • ABX error rate on embeddings

    • Scale is \([0, 1]\), lower is better

  • Lexical and Syntactic

    • Mean correct / incorrect classification accurary

    • Scale is \([0, 1]\), higher is better

  • Semantic

    • Human judgement correlation coeficient (x 100)

    • Scale is \([-100, 100]\), far from 0 is better

Phonetic (Within) Phonetic (Across) Lexical Syntactic Semantic
# Author Budget Set clean other clean other synth. libri.