INDEX
Explanations
temperature
This neuron activates on words related to temperature—especially those containing the substring “therm” (e.g., temperature, thermal, thermocouple, isothermal, etc.).
New Auto-Interp
Negative Logits
이크
-0.07
FLT
-0.07
点
-0.06
resilience
-0.06
looping
-0.06
-sample
-0.06
coerc
-0.06
هفته
-0.06
ندر
-0.06
dark
-0.06
POSITIVE LOGITS
"@
0.08
?>"><
0.07
"></
0.07
_front
0.07
furnished
0.07
Obama
0.06
oundation
0.06
Investigation
0.06
sure
0.06
swagen
0.06
Activations Density 0.006%