INDEX
Explanations
This neuron activates on closing parentheses that end parenthetical citations or numeric references (i.e., the “)” at the end of parenthetical remarks).
New Auto-Interp
Negative Logits
Tech
-0.07
TZ
-0.06
reinc
-0.06
rench
-0.06
niž
-0.06
Moon
-0.06
beach
-0.06
печ
-0.06
trx
-0.06
EMALE
-0.06
POSITIVE LOGITS
**)
0.09
.DriverManager
0.09
:')
0.08
.)
0.08
!)
0.08
")
0.08
.")
0.08
»)
0.08
)
0.08
.").
0.08
Activations Density 0.118%