"The highlighted tokens are predominantly Turkish morphemes, suffixes, and roots that form or modify nouns, verbs, and adjectives, often marking grammatical features such as possession, plurality, tense, or case. Many activations correspond to common Turkish word endings or inflectional affixes, as well as roots of frequently used words, indicating a focus on morphological structure and word formation in Turkish text."
Score Type | Accuracy | Precision | Recall | F1 score | TPR | TNR | FPR | FNR |
---|---|---|---|---|---|---|---|---|
detection | 0.84 | 1.0 | 0.68 | 0.81 | 0.68 | 1.0 | 0.0 | 0.32 |
fuzz | 0.85 | 1.0 | 0.7 | 0.824 | 0.7 | 1.0 | 0.0 | 0.3 |