INDEX
Explanations
body parts
The neuron detects tokens used to label sensory attributes and body parts (e.g. “body,” “parts,” and slash-prefixed tags like “/smell,” “/touch,” “/texture,” etc.).
New Auto-Interp
Negative Logits
dados
-0.07
-0.06
була
-0.06
размещ
-0.06
conclusion
-0.06
Rae
-0.06
была
-0.06
=s
-0.06
eckého
-0.06
arrera
-0.06
POSITIVE LOGITS
.edu
0.07
"struct
0.07
.TypeOf
0.06
\<^
0.06
thuyết
0.06
Prof
0.06
FloatingActionButton
0.06
/exp
0.06
.intent
0.06
�ん
0.06
Activations Density 0.005%