INDEX
Explanations
instances of the letter 'h', particularly in various forms and contexts
New Auto-Interp
Negative Logits
469
-0.16
468
-0.16
wers
-0.16
loh
-0.15
rite
-0.15
iese
-0.15
476
-0.15
ège
-0.15
RITE
-0.14
lesen
-0.14
POSITIVE LOGITS
ound
0.32
ater
0.28
ate
0.28
OUND
0.27
ounds
0.26
ating
0.26
ulk
0.25
obo
0.25
ates
0.25
atchet
0.24
Activations Density 0.035%