INDEX
Explanations
depending on context or unspecified
The neuron detects salient content-bearing tokens—important nouns, headings, or proper nouns (key words that carry the document's main content).
New Auto-Interp
Negative Logits
perfettamente
0.35
coro
0.35
ocasiones
0.34
ottim
0.34
quitar
0.33
fürs
0.33
赚钱
0.33
これで
0.32
kullanılır
0.32
efficiently
0.32
POSITIVE LOGITS
unspecified
1.05
undetermined
0.88
unknown
0.86
depending
0.85
unidentified
0.82
неизвест
0.82
depending
0.79
undisclosed
0.78
未知
0.78
unknown
0.77
Activations Density 0.984%