INDEX
Explanations
The neuron responds primarily to the verb “crawl,” especially when it’s used to describe crawling away or hiding.
New Auto-Interp
Negative Logits
9
-0.07
цер
-0.07
undecided
-0.07
Mixing
-0.07
909
-0.06
september
-0.06
Terminator
-0.06
.IS
-0.06
Mark
-0.06
объ
-0.06
POSITIVE LOGITS
crawl
0.14
crawled
0.11
crawling
0.10
Craw
0.10
Crawford
0.09
crawl
0.09
rawl
0.09
crawler
0.08
�
0.08
crawler
0.07
Activations Density 0.006%