INDEX
Explanations
Informal conversations
The neuron flags tokens that occur in the actual dialogue lines (character speech) as opposed to narrative, stage directions, or metadata.
New Auto-Interp
Negative Logits
tròn
-0.07
تی
-0.07
質
-0.06
برای
-0.06
BY
-0.06
گه
-0.06
printer
-0.06
tl
-0.06
meniz
-0.06
bucket
-0.06
POSITIVE LOGITS
Decision
0.07
parameters
0.07
กระ
0.06
regation
0.06
lobe
0.06
ｯ
0.06
reinc
0.06
() ↵
0.06
GER
0.06
opes
0.06
Activations Density 0.206%