INDEX
Explanations
ubiquitous/omnipresent
The neuron activates on words that convey high prevalence or ubiquity—i.e., adjectives meaning “widely present” or “very common.”
New Auto-Interp
Negative Logits
/values
-0.07
ΙΟ
-0.06
立て
-0.06
新闻
-0.06
Fel
-0.06
طل
-0.06
cle
-0.06
Scha
-0.06
폭
-0.06
Tanks
-0.06
POSITIVE LOGITS
ubiquitous
0.10
pervasive
0.08
Spread
0.07
occupying
0.07
ubiqu
0.06
statewide
0.06
overarching
0.06
Omni
0.06
äch
0.06
-present
0.06
Activations Density 0.007%