INDEX
Explanations
Prior knowledge or experience
The neuron fires on direct second-person “you” introductions (e.g. “You’ve heard of…,” “You may have seen…”)—i.e. conversational reader-address phrases with “you” plus a perfect/modal construction.
New Auto-Interp
Negative Logits
(pk
-0.06
Rebellion
-0.06
Sodium
-0.06
ception
-0.06
_requested
-0.06
pyt
-0.06
-save
-0.06
memcpy
-0.06
_tuples
-0.06
buffers
-0.06
POSITIVE LOGITS
腐
0.07
/--
0.07
excuse
0.06
suoi
0.06
noe
0.06
aware
0.06
Asheville
0.06
-equiv
0.06
湖
0.06
Fot
0.06
Activations Density 0.022%