INDEX
Explanations
references to durable and heat-resistant materials
New Auto-Interp
Negative Logits
iola
-0.20
è¡£
-0.15
alam
-0.15
arih
-0.15
_fence
-0.15
ÑĤÑİ
-0.14
æķħ
-0.14
ingu
-0.14
apan
-0.14
æ²¹
-0.14
POSITIVE LOGITS
insulated
0.29
insulation
0.27
straw
0.27
leak
0.26
lid
0.26
cup
0.25
Leak
0.25
sip
0.25
drinking
0.25
cups
0.24
Activations Density 0.015%