INDEX
Explanations
nouns and phrases related to dog breeds and training
New Auto-Interp
Negative Logits
hea
-0.17
oose
-0.16
rella
-0.14
kov
-0.14
Wie
-0.14
byn
-0.14
merch
-0.13
leck
-0.13
Trace
-0.13
lookahead
-0.13
POSITIVE LOGITS
447
0.16
Clement
0.15
hog
0.15
warts
0.14
849
0.14
wargs
0.14
Translation
0.14
à¥ĩद
0.14
habit
0.14
Hugo
0.14
Activations Density 0.804%