INDEX
Explanations
magic, spells
This neuron activates on mentions of “magic” (especially paired with “system”) and related instructional context, i.e. when the text is describing or outlining a fantasy magic system.
New Auto-Interp
Negative Logits
.Gray
-0.07
jente
-0.07
initializes
-0.07
(headers
-0.06
cretion
-0.06
etes
-0.06
мага
-0.06
$criteria
-0.06
旦
-0.06
gems
-0.06
POSITIVE LOGITS
_EM
0.07
Look
0.07
Independ
0.07
LCD
0.07
INCIDENTAL
0.06
дослід
0.06
Americans
0.06
Retail
0.06
Discover
0.06
Instructor
0.06
Activations Density 0.031%