INDEX

Explanations

continue reading

The neuron fires on content‐break or “read more” markers (e.g. “Continue,” “Continue reading,” “[...],” “…”).

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

-1.98

-1.73

 saranno

-1.44

 разом

-1.43

上げる

-1.41

 with

-1.36

 will

-1.35

 only

-1.34

 чесно

-1.32

be

-1.30

POSITIVE LOGITS

 estekak

1.86

ݯ

1.77



1.66

ー

1.63

aneously

1.62



1.57

hè

1.55

})

1.52

[]

1.52

 onError

1.51

Activations Density 0.002%