INDEX
Explanations
sentence-ending punctuation, indicating the conclusion of thoughts
New Auto-Interp
Negative Logits
FetchType
-0.84
odotus
-0.75
конец
-0.75
ugeot
-0.74
vixion
-0.72
arangay
-0.71
NUMX
-0.69
LEncoder
-0.68
ésult
-0.68
-0.67
POSITIVE LOGITS
.
1.15
))).
1.09
<eos>
1.08
}$.
1.07
$.
1.05
).
1.03
}.
1.03
])).
1.03
↵↵
1.03
))$.
1.03
Activations Density 2.565%