INDEX
Explanations
Decision
The neuron looks for the word “Decision” in court‐case citation lines.
New Auto-Interp
Negative Logits
Gund
-0.07
meye
-0.07
ecure
-0.07
Над
-0.07
Conversion
-0.06
_CHANNELS
-0.06
Moh
-0.06
KN
-0.06
MOZ
-0.06
Oxford
-0.06
POSITIVE LOGITS
DELAY
0.07
click
0.07
suppose
0.07
dodge
0.07
zin
0.06
.tell
0.06
amız
0.06
yes
0.06
tweaked
0.06
placeholder
0.06
Activations Density 0.002%