INDEX
Explanations
operation
This neuron fires on occurrences of the French word “opération” (and its inflections) in the text.
New Auto-Interp
Negative Logits
red
-0.07
presenter
-0.07
Red
-0.06
价格
-0.06
manuscript
-0.06
STAR
-0.06
двер
-0.06
_APPEND
-0.06
zk
-0.06
patch
-0.06
POSITIVE LOGITS
operations
0.09
Operation
0.08
_operation
0.08
(operation
0.08
操作
0.08
功
0.08
operation
0.08
Operation
0.08
mic
0.08
lac
0.07
Activations Density 0.019%