"The highlighted tokens are predominantly prefixes, roots, or short morphemes within words, often marking the beginning or core of nouns, verbs, or adjectives across multiple languages. These segments frequently correspond to meaningful subword units, such as roots, affixes, or syllables, and are often found in proper names, technical terms, or compound words. The pattern reflects a focus on linguistically significant subword structures that contribute to word formation and meaning."
Score Type | Accuracy | Precision | Recall | F1 score | TPR | TNR | FPR | FNR |
---|---|---|---|---|---|---|---|---|
detection | 0.54 | 0.525 | 0.84 | 0.646 | 0.84 | 0.24 | 0.76 | 0.16 |
fuzz | 0.55 | 0.527 | 0.98 | 0.685 | 0.98 | 0.12 | 0.88 | 0.02 |