INDEX
Explanations
questions and answers
tokens marking infinitival or prepositional relations (especially the word "to" and similar short function words like "in") in questions and requests.
New Auto-Interp
Negative Logits
рог
-0.06
?
-0.06
novation
-0.06
quals
-0.06
ník
-0.06
вав
-0.06
answer
-0.06
节
-0.06
genome
-0.06
Strip
-0.06
POSITIVE LOGITS
.gson
0.07
(FLAGS
0.06
důležit
0.06
WELL
0.06
compra
0.06
.CH
0.06
━━━━━━━━
0.06
эту
0.05
_rules
0.05
vra
0.05
Activations Density 0.357%