"The highlighted tokens are predominantly morphemes, suffixes, prefixes, or short stems within words across multiple languages, often marking grammatical features such as tense, case, plurality, or forming part of proper nouns and place names. These segments are crucial for word formation and meaning, and their activation suggests a focus on subword units that carry significant syntactic or semantic information."
Score Type | Accuracy | Precision | Recall | F1 score | TPR | TNR | FPR | FNR |
---|---|---|---|---|---|---|---|---|
detection | 0.53 | 0.515 | 1.0 | 0.68 | 1.0 | 0.06 | 0.94 | 0.0 |
fuzz | 0.53 | 0.515 | 1.0 | 0.68 | 1.0 | 0.06 | 0.94 | 0.0 |