INDEX
Explanations
months and topics
This neuron detects salient topical content words—domain-specific nouns and named entities that carry the main subject matter of a passage.
New Auto-Interp
Negative Logits
інки
0.65
ਦਵਾਈ
0.64
не
0.63
}$;
0.63
ਵਿੱਚ
0.60
ਾਸ
0.60
с
0.60
нов
0.59
maladies
0.59
ENCIAS
0.59
POSITIVE LOGITS
September
0.70
March
0.68
November
0.65
June
0.63
April
0.60
Saturday
0.60
Thursday
0.59
frac
0.59
August
0.57
October
0.57
Activations Density 0.028%