INDEX

Explanations

conditional suggestions to go or visit

The neuron flags strong recommendation or call-to-action language—words and phrases urging you to “must,” “go,” “try,” “visit,” “don’t miss,” etc.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

quela

-1.25

 that

-1.16

our

-1.13

we

-1.06

 thanks

-1.00

parently

-0.97

 препят

-0.96

 continuare

-0.96

 continue

-0.95

tionally

-0.95

POSITIVE LOGITS

if

1.80

 хотя

1.73

 если

1.65

至少

1.63

 yourself

1.63

 atleast

1.41

如果

1.38

 nếu

1.37

如果你

1.33

 almeno

1.33

Activations Density 0.021%