INDEX
Explanations
phrases indicating sources of information or funding
New Auto-Interp
Negative Logits
lah
-0.19
ests
-0.18
tn
-0.17
ly
-0.16
ew
-0.16
ed
-0.16
rello
-0.15
hev
-0.15
utures
-0.15
isser
-0.14
POSITIVE LOGITS
forge
0.38
book
0.32
material
0.31
code
0.30
books
0.28
.unsplash
0.27
materials
0.26
-code
0.26
Material
0.26
material
0.25
Activations Density 0.040%