INDEX
Explanations
explaining criteria and situations
textual structure and punctuation cues—especially contractions, hyphenated phrases, list/section introducers (like colons), and numeric identifiers—indicating formatting or metadata rather than core content.
New Auto-Interp
Negative Logits
\
0.22
to
0.22
ப்ப
0.21
о
0.20
อ
0.20
ERS
0.20
hydroelectric
0.20
в
0.19
С
0.19
с
0.19
POSITIVE LOGITS
faptul
0.24
ነገር
0.23
unequiv
0.22
natthi
0.21
எதுவும்
0.21
experiencias
0.21
situación
0.21
conteú
0.21
بابەت
0.21
whether
0.21
Activations Density 2.777%