INDEX

Explanations

quite brave some forms

The neuron responds to exclamation-point punctuation marks.

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

PAIGN

-1.41

 about

-1.34

 prace

-1.32

 не

-1.30

 desear

-1.28

arskie

-1.23

 там

-1.22

rei

-1.22

ممنون

-1.22

 puis

-1.21

POSITIVE LOGITS

</h2>

1.80

 from

1.76

 should

1.69

 itself

1.63

</h1>

1.62

 could

1.57

 would

1.49

 while

1.41

年来

1.38

 solchen

1.32

Activations Density 0.268%